| Data source | ||
|---|---|---|
| Response or edited | Imputed | |
| % | ||
| Sales of goods manufactured | 83.0 | 17.0 |
| Raw materials and components | 76.1 | 23.9 |
| Goods / work in process | 79.2 | 20.8 |
| Finished goods manufactured | 76.3 | 23.7 |
| Unfilled Orders | 85.1 | 14.9 |
| Capacity utilization rates | 62.7 | 37.3 |
Date: July 2022
Program manager: Director, Centre for Population Health Data
Director General, Health, Justice, Diversity and Inclusion
Personal information collected through the voluntary Canadian Health Measures Survey (CHMS) is described in Statistics Canada's "Health Surveys" Personal Information Bank. The Personal Information Bank refers to information collected through Statistics Canada's health surveys and includes a variety of topics and information such as: name, contact information, citizenship status, education, employment, financial, language, health and medical information (from blood, urine samples), place of birth, pregnancy, breastfeeding, sleep habits, sexual behavior, nutrition, alcohol, e-cigarette/cigarette use, medication/drug use, physical attributes, physical activity, neighborhood environment, health, COVID-19, and medical information using blood, saliva, and urine samples.
The "Health Surveys" PIB (Bank number: StatCan PPU 806) is published on the Statistics Canada website under the latest Information about programs and Information Holdings chapter.
Statistics Canada, under the authority of the Statistics ActFootnote 1, conducts the voluntary Canadian Health Measures Survey (CHMS). The CHMS aims to collect important health information through a household interview and direct physical measures at a mobile examination centre (MEC), sometimes referred to as a mobile clinic. The main objective of the survey is to gather information that helps improve the prevention, diagnosis and treatment of illnesses, as well as promote the health and wellness of Canadians. In addition, this survey helps shed light on illnesses and reveal the extent to which many diseases may be undiagnosed among Canadians, and enables health professionals and researchers to address public health challenges.
Fundamentally, the value of the data lies in analysis and interpretation. Demographic, socio-economic, and health characteristics such as age, sex and gender, marital status, ethnic origin, eating habits, physical activity and chronic diseases become especially meaningful when analyzed in relation to one another.
Aggregate survey results are published in the Daily (the Agency's official release bulletin) summarizing the survey findings in the form of profiles and cross-tabulations, anonymized public use microdata files (PUMFs)Footnote 2, and analytical reports. These data are fully anonymized and non-confidential, without any direct personal identifiers, which prevents the possibility of identifying individuals. Through a data sharing agreement, and as permitted by section 17(2) of the Statistics Act, Health Canada (HC) and the Public Health Agency of Canada (PHAC) have access to the data, with all personal identifiers removed, in the Research Data CentresFootnote 3 and are only permitted to release aggregate results, which are fully anonymized and non-confidential. They will use these data to: establish Canadian health reference ranges, identify the prevalence of health indicators (e.g., diabetes) and monitor trends, examine relationships between health indicators (e.g., blood pressure, cardiovascular disease), and provide insight on fitness and health.
To ensure the continued relevance of the CHMS, Statistics Canada conducts formal consultations at the start of each new survey cycle. Data users and stakeholders are invited to provide feedback on what information they use, for what purpose and what, if any, data gaps Statistics Canada should consider addressing in the next cycle. The most recent consultation has led to the following changes to the seventh cycle of data collection which is scheduled for 2022:
The CHMS survey sample is being extended to include 1 to 2 year-olds. The legal parent or guardian responds to the CHMS household interview on behalf of the child, and must be present with the child at the MEC and provide written consent for the child to participate in the tests which include height and weight, bone health via DXA and blood collection.
Body composition and bone health will now be evaluated by anthropometry and dual energy X-ray absorptiometry (DXA) scans of the whole body, femur, and anterior-posterior (AP) lumbar spine.
An oral health component was included in CHMS Cycle 1 (2007-2009); it was planned at that time to conduct the same oral health measurements every fifteen years. This component is essential in identifying oral health trends in Canada and provide evidence to support informed policy and program development at all levels of government.
The current device (Actical) used to collect physical activity measures is not well-suited for the collection of sleep data, therefore a new device (Actigraph) will be used in this cycle. The collection of sleep data will establish 24-hour movement behaviours and their association with health conditions and risk factors.
The device used to measure blood pressure in Cycles 1 to 6 of the survey was provided by BPTru™ which has since ceased its operations. Therefore, a new device (OMRON) which is validated and highly recommended by experts will be used in this cycle.
In order to ensure a safe work environment for the MEC staff, protect the health of respondents, and mitigate the risk of SARS-CoV-2 transmission during collection, the CHMS will include a COVID-19 screening strategy for both MEC staff, respondents, and any other person who could be required to access the MEC (e.g., contractor, family member).
All other personal information collected through the survey remains the same as in previous cycles and is described in previous Canadian Health Measures Survey PIAsFootnote 4 as well as in the Generic PIA for Statistics Canada's Statistical Programs. Health Canada (HC) and the Public Health Agency of Canada's (PHAC) Research Ethics Board have reviewed Cycle 7 content.
While the Generic Privacy Impact Assessment (PIA) addresses most of the privacy and security risks related to statistical activities conducted by Statistics Canada, this supplement describes any potential new risks associated with the changes introduced in this survey cycle. This supplement also presents an analysis of the necessity and proportionality of these new elements. As is the case with all PIAs, Statistics Canada's privacy framework ensures that elements of privacy protection and privacy controls are documented and applied.
The changes in the collection and use of personal information for Cycle 7 of the Canadian Health Measures Survey can be justified against Statistics Canada's Necessity and Proportionality Framework:
The personal information collected for this age group will provide reliable information on the health and wellness of young Canadians. There is substantial interest for information on this age group as there is currently limited data available for this population in Canada. Dual Energy X-ray Absorptiometry (DXA) will be conducted in partnership with the Children's Hospital of Eastern Ontario (CHEO) and the environmental measures will be analyzed with the support of Health Canada. Exposure to environmental chemicals during these critical development years can have a lasting impact on the child's future health outcomesFootnote 5.
The personal identifiers for the children included in the survey, as well as for all other respondents (full name, date of birth, and provincial or territorial health card number) are collected for linkage purposes. Personal identifiers are removed from the data file and stored separately and securely, once the linkage has been performed. Statistics Canada's microdata linkage and related statistical activities were assessed in Statistics Canada's Generic Privacy Impact AssessmentFootnote 6. All data linkage activities are subject to established governanceFootnote 7, and are assessed against the privacy principles of necessity and proportionalityFootnote 8. All approved linkages are published on Statistics Canada's websiteFootnote 9.
The addition of this age group was submitted for review to the Public Health Agency of Canada-Health Canada Research Ethics Board to solicit their health and medical expertise in order to ensure ethical issues were considered, to ensure that internationally recognized ethical standards for human research were met and maintained, and to ensure minimal respondent burden while maximizing the data potential.
The information collected with the dual energy X-ray absorptiometry (DXA) will fill a gap in quality national estimates of body composition and bone density. It will help address a wide range of priority policy questions pertaining to bone health which cannot currently be addressed with accuracy. The DXA will allow the investigation of health-related topics including osteoporosis, obesity and cardiovascular disease risk.
Body composition and bone health will be evaluated using anthropometry and DXA scans of the whole body, femur, and anterior-posterior (AP) lumbar spine. The information provided by the scans will allow: nationally representative data on total and regional bone mineral content, lean mass, fat mass, and percent fat overall and for age, gender, and racial/ethnic groups; estimates of obesity, defined as an excess of body fat; data to study the association between body composition and other health conditions and risk factors, such as cardiovascular disease, diabetes, hypertension, physical activity, and dietary patterns; estimates of the prevalence of osteoporosis and low bone mass; and the first estimates of the prevalence of vertebral fractures and abdominal aortic calcification.
The oral health component is Canada's only complete overview of the current oral health status of Canadians. The Canadian Health Measures Survey (CHMS) is the best survey to conduct data collection for these measures as it combines both clinical measurements and a household questionnaire.
The objectives of the oral health component are to evaluate the association of oral health with major health concerns such as diabetes, respiratory and cardiovascular diseases. It will help determine relationships between oral health and certain risk factors such as poor nutrition, environmental factors such as levels of fluorides in water, and socioeconomic factors related to low income levels and education.
In order to remain relevant and meet current and foreseeable future research needs, Statistics Canada will measure all respondent activities within a 24-hour period, including sleep, as opposed to only waking hours. The current device (Actical) is not well-suited for the collection of sleep data, therefore Statistics Canada will be using the Actigraph wGT3X-BT for use in future cycles of the survey.
This new monitor captures and records high resolution, raw acceleration data, which is converted into a variety of objective activity and sleep measures using publicly available algorithms developed and validated by members of the academic research community. The high quality, timely and relevant data collected will determine the degree to which the Canadian Physical Activity GuidelinesFootnote 10 are being met and allow to measures the health benefits and potential risk of developing diseases. Available historical data will also allow for the monitoring of trends.
The Canadian 24-hour Movement Guidelines, created by the Canadian Society for Exercise Physiology (CSEP), provide guidance on the optimal amount of physical activity, sedentary behavior, and sleep, and the best combination of these behaviours, for Canadians of all ages. Evidence suggests that people who follow these guidelines can improve their health. Benefits include lower risk of cardiovascular disease, obesity, and cancer, improved bone health, and enhanced psychosocial health.
Currently, there is an information gap at the national level concerning the percentage of the Canadian population that meets these guidelines. In partnership with Health Canada and the Public Health Agency of Canada, Statistics Canada will use the information collected from the CHMS to evaluate the 24-hour movement behaviours and their association with health conditions and risk factors.
During Cycle 6 collection of the CHMS, the company BPTru™ which provided the device for the collection of blood pressure measures ceased its operations.
To replace the device, Statistics Canada has opted for the OMRON HEM-907-XL which is highly recommended by experts, and is used for other national studies and major trials such as the National Health and Nutrition Examination Survey and Systolic Blood Pressure Intervention Trial (Center for Disease Control and Prevention, USA). It was also validated by the Association for the Advancement of Medical Instrumentation and the International Protocol of the European Society of HypertensionFootnote 11.
In order to ensure a safe work environment for the MEC staff, protect the health of respondents, and mitigate the risk of SARS-CoV-2 transmission during collection, Statistics Canada is planning on implementing a COVID-19 screening strategy for both MEC staff and respondents.
During the MEC appointment reminder phone call (24-48 hours before the appointment), the Centre Coordinator will confirm verbally with the respondent that they are not symptomatic, and understands that they will have to undergo COVID-19 screening measures upon their arrival at the MEC to participate in the MEC visit. If the respondent is symptomatic or refuses to follow the protocols their appointment will be cancelled.
Upon arrival, every individual entering the MEC will be required to follow the COVID-19 screening protocols established in collaboration with Statistics Canada Occupational Health and Safety group. They will be provided with a consent form (Appendix A) for the COVID-19 Screening Protocols, which will consist of a screening questionnaire (Appendix B) and temperature check (not recorded). The form and questionnaire will have the name of the individual, the date and the CLINIC ID of the respondent and will be stored in a secured drawer in the administrative room in the MEC. During the day, the secured drawer is accessible using a key by all deemed Statistics Canada MEC employees (approximately 20) for them to put the respondents file away after the appointment. The drawer is locked when the MEC is closed and the Site Manager, Team Leaders (2) and Clinic Coordinators (2) have access to the key. All respondent files are shipped to Head Office at the end of every MEC collection site (MEC changes site every 5-6 weeks) in a Versapack (same shipping procedure as previous cycles; the Versapack is the secure method used by Statistics Canada to ship confidential packages from the MEC to Statistics Canada's Head Office). At Head Office they are stored in a locked filing cabinet until they can be sent to Statistics Canada's Operations Integration team to be scanned and stored electronically. The documents will be stored and retained/destroyed as per Statistics Canada's retention standards. In this case, COVID-19 screening material is considered "work files in support of design and collection", making them transitory files with a retention requirement described as, "Delete as soon as they have served their use or purpose, within a year of the end of collection." with discretion also given to the responsible manager according to the following criteria:
Respondents who do not consent to COVID-19 screening will be informed that they are ineligible to participate in the direct measures component of the survey.
The COVID-19 Screening Strategy will be subject to change depending on the evolution of the pandemic, recommendations from the Occupational Health and Safety group (that will be consulted regularly through collection) and provincial public health guidelines. Should any of those changes have a privacy impact, an amendment to this Sought PIA will be created.
"In determining the appropriate level of additional documentation required, the responsible manager's decision should be based on the use of statistical microdata files and their expected retention. The following are suggested criteria that could be considered in making this determination:
Use:
Retention:
The CHMS is carefully designed to produce relevant, high priority, statistically meaningful information. Although a great deal of health knowledge can be obtained through survey interviews and administrative databases, the most accurate information that is relevant to the present and future health of the population can only be obtained through direct measures of physical characteristics.
A sample size of 670 1-2-year-olds has been assessed as necessary by methodologists to produce statistics of sufficient quality for that age group. The total sample size for the survey is therefore consequentially increasing from 5700 to 6370. Data analysis and reporting will not be modified and will follow approved guidelines used in previous cycles.
Dual-energy X-ray absorptiometry (DXA) is a clinically proven, accurate, and reproducible method of measuring bone mineral density in the lumbar spine, proximal femur, and whole body. Body composition and bone health will be evaluated starting in Cycle 7 of the CHMS by anthropometry and DXA scans of the whole body, femur, and anterior-posterior (AP) lumbar spine.
All respondents will be asked to undergo scans. The final report will only summarize group results, and no individual, personal or confidential information will be shared.
The oral health component is Canada's only complete overview of the current oral health status of Canadians. The Canadian Health Measures Survey (CHMS) is the best survey to conduct data collection for these measures as it combines both clinical measurements and a household questionnaire.
The Actigraph will measure raw acceleration, activity counts, energy expenditure, MET rates, steps taken, physical activity intensity, activity bouts, sedentary bouts, body position, sleep latency, total sleep time, wake after sleep onset, sleep efficiency, and ambient light. Raw data will be stored using participant numbers as identifiers.
The final report will only summarize group results, and no individual, personal or confidential information will be shared.
The decision was made to use the OMRON HEM-907-XL because it was highly recommended by experts and it is used in other national studies and major trials such as the National Health and Nutrition Examination Survey and Systolic Blood Pressure Intervention Trial (Center for Disease Control and Prevention, USA). It has also been validated by the Association for the Advancement of Medical Instrumentation and the International Protocol of the European Society of HypertensionFootnote 12.
The final report will only summarize group results, and no individual, personal or confidential information will be shared.
COVID-19 screening strategies will evolve through survey collection to ensure they are following the provincial and federal health & safety guidelines at the minimum. They will also be reviewed by Statistics Canada's Occupational Health and Safety group to ensure that all necessary precautions to protect employees and survey respondents are taken. COVID-19 screening results will not be retained nor shared.
The data collected will contain only the variables required to achieve the statistical goals of the CHMS. Statistics Canada directives and policies with respect to data collection and publication will be followed to ensure the confidentiality of the data. Individual responses will be grouped with those of others when reporting results. Individual responses and results for very small groups will never be shared with government departments or agencies.
The benefits of the findings, which are expected to support services aimed at improving the prevention, diagnosis and treatment of illnesses and to promote the health and wellness of Canadians, are believed to be proportional to the potential risks to privacy.
The CHMS is one of the only sources of information for small geographic areas, based on the same statistical concepts for the entire country, and the only source of information for many health characteristics.
Data linkages are used between CHMS data and other sources of information for statistical analyses; to evaluate data quality, to assist with data processing, and for direct replacement of data when the quality is deemed appropriate.
Some questions and measurements contained in the CHMS are considered sensitive as they relate to an individual's health and wellness; however, the overall risk of harm to the survey respondents has been deemed manageable with existing Statistics Canada safeguards that are described in Statistics Canada's Generic Privacy Impact Assessment, including, but not limited to the following:
For the household interview portion of the CHMS, participants will be informed in the invitation letter that their participation is voluntary and before being asked any questions. For the direct physical measures element, participants will also be informed of the voluntary nature of the CHMS and the topic of each component before participating. Finally, prior to entering the MEC, participants will be informed that the COVID-19 screening is voluntary, yet is necessary to proceed with the clinical component of the survey.
Variables that directly identify respondents will be separated from the data files in the first stage of data processing and placed in a secure location with controlled access. Variables that might indirectly identify respondents are examined and modified as necessary in order to protect the privacy and confidentiality of respondents. Individual responses will be grouped with those of others when reporting results. Individual responses and results for very small groups will never be published or shared with government departments or agencies. Careful analysis of the data will be performed prior to the publication and sharing of aggregate data to ensure that marginalized and vulnerable communities are not disproportionally impacted.
The linkage of CHMS data with other sources of information will be used in statistical studies to evaluate data quality and the impact of non-response, to improve and assist with data editing and imputation, and for direct replacement of data in presence of partial non-response when the quality is deemed appropriate. The linkage files will be used only within Statistics Canada for methodological research, development and processing.
Security measures for linkage keys and administrative files respect the policies, directives and guidelines for information technology security at Statistics Canada. When linkage is required, it is done using anonymized statistical identifiers ("linkage keys") and, as a result, no linked file contains personal identifiers such as name, phone number and address (excluding postal code). These anonymized statistical identifiers are used to link to other sources of information for statistical purposes only. The personal identifiers obtained are removed from the rest of the information and securely stored with restricted access with an approved operational requirement to access them, and whose access is removed when no longer required.
It is the policy of Statistics Canada to provide all respondents with information about: the purpose of a survey (including the expected uses and users of the statistics to be produced from the survey), the authority under which the survey is taken, the mandatory or voluntary nature of the survey, confidentiality protection, the record linkage plans and the identity of the parties to any agreements for sharing of the information provided by those respondents, where applicable.
For Cycle 7 of the CHMS, this information is provided in the letter of invitation, in the survey's consent form, and in Frequently Asked Questions (FAQs) accessible through the CHMS websiteFootnote 13.
This assessment concludes that, with the existing Statistics Canada safeguards, any remaining risks are such that Statistics Canada is prepared to accept and manage the risk.
This Supplementary Privacy Impact Assessment has been reviewed and recommended for approval by Statistics Canada's Chief Privacy Officer, Director General for Modern Statistical Methods and Data Science, and Assistant Chief Statistician for the Enterprise Statistics Field.
The Chief Statistician of Canada has the authority for section 10 of the Privacy Act for Statistics Canada and is responsible for the Agency's operations, including the program area mentioned in this Supplementary Privacy Impact Assessment.
This Privacy Impact Assessment has been approved by the Chief Statistician of Canada.
| Month | Sales of goods manufactured | Raw materials and components inventories | Goods / work in process inventories | Finished goods manufactured inventories | Unfilled Orders |
|---|---|---|---|---|---|
| % | |||||
| August 2021 | 0.74 | 1.04 | 1.53 | 1.81 | 1.50 |
| September 2021 | 0.79 | 1.03 | 1.54 | 1.83 | 1.41 |
| October 2021 | 0.76 | 1.03 | 1.52 | 1.73 | 1.46 |
| November 2021 | 0.73 | 1.00 | 1.62 | 1.57 | 1.34 |
| December 2021 | 0.75 | 1.01 | 1.81 | 1.56 | 1.46 |
| January 2022 | 0.78 | 1.12 | 1.82 | 1.85 | 1.43 |
| February 2022 | 0.73 | 1.14 | 1.64 | 1.77 | 1.38 |
| March 2022 | 0.71 | 1.13 | 1.52 | 1.66 | 1.44 |
| April 2022 | 0.69 | 1.19 | 1.51 | 1.62 | 1.49 |
| May 2022 | 0.67 | 1.16 | 1.54 | 1.68 | 1.41 |
| June 2022 | 0.69 | 1.15 | 1.55 | 1.76 | 1.44 |
| July 2022 | 0.70 | 1.11 | 1.68 | 1.49 | 1.36 |
| August 2022 | 0.69 | 1.13 | 1.79 | 1.57 | 1.39 |
Release date: October 17, 2022
Catalogue number: 45200003
ISSN: 2816-2250
Listen to "Eh Sayers" on:
StatisticsCanada
@statcan_eng
@StatCan_eng
StatCanada
StatisticsCanadaPoverty graphic (JPG, 1.84 MB)
It used to be that Statistics Canada didn't measure poverty. Not exactly. Poverty is complex, and there wasn't a single definition that everyone agreed on. So while StatCan did measure low income and other income inequality indicators, it didn't measure poverty per se. That is, until 2018, when the government chose to use the Market Basket Measure, or MBM, as Canada's Official Poverty Line. That means that the government now uses the MBM to track its poverty reduction targets.
But something strange happened during the pandemic: in 2020 the poverty rate fell. And it fell quite a bit. In fact, the poverty rate dropped in one year almost as much as it had in the four preceding years.
But why? What happened? Will the poverty rate continue to fall? And what happens if it hits zero? How would health outcomes change? Education outcomes? People's general happiness and well-being?
Has there ever been a time and place in Canada where the poverty rate was zero? The closest may be the Mincome Experiment of the 1970s in Manitoba. Many Canadians have never heard of this guaranteed income experiment, but it offers a glimpse at what eliminating poverty might look like.
To learn more we spoke with Burton Gustajtis, an economist from Statistics Canada, Evelyn Forget, a Professor of Economics and Community Health Sciences at the University of Manitoba and Kevin Milligan, a Professor of Economics in the Vancouver School of Economics at the University of British Columbia.
Tegan Bridge
Burton Gustajtis, Evelyn Forget, and Kevin Milligan
Volume
Current position:00:00:00
Total time:--:--:--
(Theme)
Tegan: Welcome to Eh Sayers, a podcast from Statistics Canada, where we meet the people behind the data and explore the stories behind the numbers. I'm your host, Tegan Bridge.
It used to be that Statistics Canada didn't measure poverty. Not exactly. Poverty is complex, and there wasn't a single definition that everyone agreed on. So while StatCan did measure low income and other income inequality indicators, it didn't measure poverty per se. That is, until 2018, when the government chose to use the Market Basket Measure, or MBM, as Canada's Official Poverty Line. That means that the government now uses the MBM to track its poverty reduction targets as well as progress made towards Canada's sustainable development goal for the elimination of poverty.
Something weird happened during the pandemic: in 2020 the poverty rate fell. And it fell quite a bit.
The poverty rate dropped to 6.4%, down from 10.3% in 2019, that's more than a third. In that single year, the rate dropped almost as much as it had in the four preceding years.
But why? What happened? Will the poverty rate continue to fall? And what happens if it hits zero? Could we be on the brink of solving poverty in Canada?
We're going to talk about what happened, but first, it would be helpful to understand what the Market Basket Measure is and how it works.
Burton: My name is Burton Gustajtis. I'm an economist at Statistics Canada.
Burton: In general, the MBM is an absolute measure of low income. It's based on the cost of a specific basket of goods and services mentor represent a modest basic standard of living for a family of four
Tegan: The MBM is like a shopping cart filled with all of the things that you need: food, clothing, footwear, shelter, transportation, and other necessities.
Burton: Each of these components, where appropriate, follow standards created by experts in their given domains. For example, the food component is based on a commonly consumed food item that represents a nutritional diet. Using the 2019 National nutritious food basket developed by Health Canada, and it's consistent with the latest Canada Food Guide.
Tegan: Experts at StatCan then look at that full shopping cart and estimate how much it costs, and that cost becomes the threshold. If you have enough disposable income to purchase that shopping cart of goods, you're living above the poverty line. If you can't, you're below it.
What about people who don't necessarily fit into a family of four structure?
Burton: For different family sizes, we use an equivalization methodology, which is an internationally recognized method of adjusting low income thresholds and income estimates for different family sizes.
Tegan: Is that like an equation?
Burton: Yes, it's called the square root equivalization methodology. So basically, the idea is that the cost for a family increase but at a decreasing rate. So the more people you get in your family, the more expensive your basket will be. But it's not at a, at a linear rate. It's at a decreasing rate.
Tegan: Gotcha. So every additional person that's not like you double the number every single time.
Burton: Yeah, exactly. That's right. Yeah. It increases, but not at a constant but constant rate like that, yeah.
Tegan: So, in addition to having some allowance for different family sizes, there's also a regional component.
Burton: This basket is costed in 53 regions across the provinces.
Tegan: Things can cost different amounts depending on where you live. The same, say, loaf of bread might have a different price if I were to buy it in Halifax, or rural Alberta, or Montréal. So the MBM takes into account the area where people live.
Does the basket change with inflation?
Burton: It does, yeah. So it's an absolute measure of poverty like I mentioned. So the contents of what that means basically is that the contents of the basket are held constant in a base year. Our current base year is 2018. And then it's adjusted annually for inflationary changes, price changes only. The contents of the basket is held constant, but the price is adjusted using the connect consumer prices index.
Tegan: The MBM thresholds are published annually. Therefore, changes in inflation you see from one month to the next are reflected in the annual updating of the MBM items' prices. We did an entire episode this past January, January 2022, about inflation and the CPI, called "Why Should You Care About Inflation?" Check that out to learn more!
Does the Market Basket measure fully capture poverty in Canada?
Burton: That's a good question. Poverty is a complicated concept. It's not just low income defined like the Market Basket Measure. It uhh, it's also, multidimensional. It's, you know, inequalities in the income distribution, being below the poverty line, entry and exit in and out of poverty. It's access to education, a well-paying jobs, social integration. You know, it's not just low income.
Tegan: The Market Basket Measure is a great tool: it's easy to understand, it takes into account differences in geography and allows for some differences in family size, and it's continually being updated, or rebased, which is the technical term, by StatCan and their partners at Employment and Social Development Canada, to ensure that it reflects the up-to-date cost of a basket of goods and services representing a modest, basic standard of living in Canada and to improve the tool and address any potential shortcomings. The Market Basket Measure is useful, but it's not the only way to track poverty. StatCan has a poverty dashboard on the website called the Dimensions of Poverty Hub with 12 additional indicators that you can check out to get a fuller picture.
Now, as I said at the top of the show, the poverty rate had been trending down before the pandemic. Between 2015 and 2019, it dropped from 14.5% to 10.3%, a difference of 4.2 percentage points over four years. What's noteworthy about 2020 is that the poverty rate dropped to 6.4%. Again, that's down from 10.3% in 2019, a difference of 3.9 percentage points, or more than a third. In that single year, the rate dropped almost as much as it had in the four preceding years.
In response to the COVID-19 pandemic, and the shutdowns and restrictions put in place to manage it, the Government of Canada introduced new income supports for individuals as well as businesses, like the Canada Emergency Response Benefit and the Canada Emergency Student Benefit.
Burton: The impact of the pandemic was not felt equally and many families did suffer. What many families do not suffer losses to employment or earnings, rather, earnings employment losses tended to be concentrated among that families and individuals on the lower, but that had lowered market income. So in response to these losses in employment and earnings a number of Canadians turned to the existing and newly announced income support measures that were put in place. These programs provided approximately $82 billion dollars in income and supported about 8.1 1,000,000 Canadian families and unattached individuals in 2020. And overall result of this was that the poverty rate fell by more than 1/3 in 2020. The decreases were universal. They were across all provinces, family types, demographic groups, although I should caution the gaps between the at risk populations and those not typically at risk of poverty remained so that although poverty decrease for everyone, the gaps between the address populations remain the same.
Tegan: It's important to note that these shifts in the poverty rate were caused by temporary government supports, so we can't expect these changes to be permanent.
We learned about the poverty rate's impressive drop, and it really got everybody on the podcast team thinking. What would Canada look like if the poverty rate hit zero? How would health outcomes change? Education outcomes? People's general happiness and well-being?
This made us really curious. Has there ever been a time and place in Canada where the poverty rate was zero? The closest may be the Mincome Experiment. We wanted to learn more, so we knew we needed to find an expert.
Evelyn: My name is Evelyn Forget. I'm a professor in the Department of Community Health Sciences at the University of Manitoba.
Tegan: Could you tell us what was the Mincome experiment?
Evelyn: the Mincome experiment happened in the mid-1970s in Canada at a time when the federal government and the various provincial governments were rethinking a lot of the social programs that were delivered in this country. And what it was guaranteed annual income experiment. That's what it was called at the time. What it meant was that everybody who participated in the experiment would receive a promise that they would receive a certain agreed upon amount of money if they had no other source of income. If they did have another source of income. If, for example, they worked and earned a little bit of money, the benefit would be reduced, but it would be reduced less than proportionately. So the guaranteed income acted both as a supplement to low wage workers and as a replacement for provincial income assistance at the time. There were two sites in Manitoba that were chosen to participate in the experiment. Winnipeg and a small town known as Dauphin Manitoba.
Tegan: Families received money for three years, from 1975 to 1978, and one of the goals was to evaluate the impact of a guaranteed annual income on the work behaviour of recipients, to test the theory that if you give people money they won't have the same incentive to work, and they'll cut their hours at work or even quit their jobs.
And what were the results of this experiment?
Evelyn: During the 1980s, there were a couple of economists at the University of Manitoba who looked at the labor market results. Derek Hum and Wayne Simpson. And they discovered what's been discovered in many other basic income and guaranteed income experiments. And that is that there really isn't much of a labor market response people who were working before the experiments started to pretty much continue to work. People who weren't working ahead of time for the most part didn't start working and didn't stop. So there. There wasn't much of a change in participation in the labor market that was caused by the basic income experiment or the guaranteed annual income experiment.
Tegan: But there were two notable exceptions to this.
Evelyn: There were, in effect, two groups of people who did work less during the experiment. One of those groups of people. Were new mothers. Umm. So if you think again back to the 1970s. Maternity leaves. Were not what they are now. There was no one year parental leave and for the most part, women were guaranteed four weeks off when they gave birth. And there were a lot of new mothers who thought that that was a rather miserly response to childbirth. And I think predictably, many of those families use the Mincome to buy themselves longer parental leaves. But the other group of people who did work less were-- and here the language turns out to be really, really important. And the language that was used in the report was young, unattached males, that is, young men who hadn't yet formed families. They weren't married. They weren't. They weren't living in, in, in committed relationships. They didn't have children, and they work less, and that seemed to feed a lot of the prejudices people had a lot of worries that people had about guaranteed income. And what I was able to do was to go back and to find some of the educational records during the period. And one of the things I showed was that there was. And nice little bubble in high school completion rates exactly during the Mincome experiment. And what that meant was that people who probably wouldn't have finished high school were able to finish high school because their families received Mincome support.
Tegan: People thought that these young men were doing as everyone feared, taking the money and running, but Evelyn's findings suggests that what might have happened instead was that young men who might otherwise have dropped out of high school to help support their families were instead given the opportunity to complete their education and graduate.
Did this experiment have any kind of impact on what kinds of work people did?
Evelyn: Well, I don't have data looking specifically at the kinds of jobs that people were doing. What I have are a lot of anecdotal reports from people who participated in the experiment. And so I was able, for example, to talk to people who use the opportunity to keep small businesses alive or to start small businesses. And I thought that was a really interesting outcome. One woman I talked to living in Dauphin. I have opened a small record shop selling, you know record players and vinyl records during the period. And she said she remembered the Mincome period as being a time when everybody had a little bit of money in their pockets. But there were a lot of stories of people using the Mincome money to invest in small businesses that were already preexisting, Umm, a lot of the people in Dauphin, for example, were farmers or related in some way to agriculture. And so Mincome stabilized their income and allowed them to invest in new equipment to build their businesses.
Tegan: So in terms of people's health, did this have any kind of impact on people's physical or mental health?
Evelyn: One of the reasons I went back looking for the Mincome records was to find out whether people's health improved. I was particularly interested in mental health, but I was interested in health and all kinds of ways. And what I was able, I was very lucky actually, because Manitoba had just moved to Universal health insurance just before the experiment began. And what I was able to do was to track down some of the participants in the Medicare Records and to look at what happened to their health. And so I was able to compare people in the experiment to a match group of people who were living in similar kinds of places the same age and sex, who didn't receive income support. And I was able to show that hospitalization rates fell pretty substantially during the experiment. Overall, the hospitalization rate fell by about 8.5%. And that's a pretty dramatic finding, a big reduction in hospitalization rates. When I looked at it a little bit more closely to find out why hospitalization rates fell, there were really two categories that stood out. The first were accidents and injuries. No, that's a big category that picks up all kinds of acute hospital admissions. You know, people have been in car accidents, accidents of all types and so on. But the other category was mental health. There was a big reduction in hospitalizations related to mental health. So that was one of the big findings, I think during this experiment it was certainly it was certainly an interesting outcome to see from a guaranteed income experiment.
I think we find similar results every time we run the similar kinds of experiments. I think that people's health inevitably improves when their income goes up. I think that's not a surprise. We see it in a lot of different kinds of programs. Umm, I think that one of the things that becomes very obvious to people is that poverty imposes a lot of costs on the economy and on society. And if you can do something that reduces the rate of poverty, you could improve living standards not only for people who are receiving the money, but for everybody who lives in a town. Everybody who lives together. Umm, so I think those things are very positive. But the basic findings I think are things we see over and over and over again. That poverty has a cost. We feel that cost and in very personal terms, in terms of our health, in terms of our wellbeing, that if you give people money for the most part, they spend it on things that improves the quality of life for themselves and their family. They invest in education, they invest in better housing, better food. So in some sense, there are no surprises there.
People who receive money at vulnerable periods in their lives can really make changes that are going to affect their health that are going to affect their lives for years and years and years to come.
Tegan: So, back to the MBM…what would happen if everyone lived above the threshold of the MBM? , by definition that would mean that there would be no poverty. Now, I'm not going to lie. We, and by 'we', I mean the podcast team who are most definitely not experts, might have gotten a bit carried away with this idea.
We asked our StatCan expert about using the Market Basket Measure in a Mincome-like way, and he very kindly explained some of the issues.
Burton: It's a statistical tool meant to be used in parallel with an income concept given the construction of its disposable income like the tenure type adjustments and the fact that, like you mentioned, the costs are not defined independently for different family sizes or constructions. It can't be used in that manner for program eligibility or sending minimum wage or for a basic income concept that it's, it's not, it's a great tool for model, for measuring poverty and income distribution, but it shouldn't be seen as a universal tool that can solve all of these problems.
Tegan: We wanted to learn more about poverty and why it's so complex, and also why we shouldn't try to use the MBM as a universal tool.
Kevin: I'm Kevin Milligan. I'm a professor of economics at the Vancouver School of Economics at the University of British Columbia.
Tegan: The market basket measure is a great reporting tool, but it reports after the fact. So , it isn't used to address need.
Kevin: So just to give an example, you know right now we have a lot of income transfers that are based on your family income and your family circumstance. Think of the Canada Child benefit, think of the GST tax credit, that lower income Canadians get and many of us got when we were students, you know, you got that check every quarter that direct deposit. Those are based on the tax filing schedule and so we just filed our taxes and April 2022. Those benefits are all being adjusted quite shortly in July 2022 for the 12 months from July 2022 to June 2023. So if you think about it, if I lost my job tomorrow, when would my Canada child benefit? When would my GSE tax credit be updated? Well, I file my taxes in April 2023, so July 2023, my check will be updated, which is maybe not great because what if my upgrade income needs are now? So that's one big challenge that you face is that cycle of how things repeated now you could argue that maybe there's ways we can do things more quickly. We could not key things off the income tax cycle and do it in some other way. But that that's the kind of challenge you have to face is how do you actually get checks into people's hands based on their current circumstance? And that's one of the many challenges that we face.
Tegan: The Market Basket Measure doesn't take into account different circumstances, like family shape and size. A family of four could mean two parents and two children, but it could also be one parent and three children. These two families would have very different needs. And that's before you add things like disability to the mix. Kevin explains.
Kevin: Our current system is really strongly based on your needs. So if you're someone who is with a disability, you're gonna have a different kind of income structure. If you're someone and even varies by disability, depends on your family circumstance, depends on a lot of different aspects of your life. So we have a whole panoply of government programs. They're pretty complicated. They often interact in poor ways, and those are not good things. But we have to understand the reason why they exist is that there are a variety of different needs. If we were to replace that whole basket, inference programs with a 'one-size fits all' program. That's kind of a flat check of some kind that doesn't depend on your needs. Then if you think about it, the people who gonna be hurt most are the people with the biggest needs because that one size fits all check is not gonna touch all the bases. That they might have in terms of their needs. And so that's a great challenge for a basic income approach is that if you try to make sure that the people with the highest needs are made whole, that they get the same kind of income transfer, you end up, kind of, well, you have to take into account all of their different kinds of disability, all of their family circumstances, all of their income patterns, and you essentially end up recreating all the complexity of the existing system. So what I'm suggesting here is no magic solution here. Sometimes. Basically income is thought of as a magic solution to things that we can wipe the table with all the complexity. And what I'm asking everyone to think about is the reason why that complexity exists is that people have complicated lives. Doesn't mean we shouldn't push back on the complexity and try to improve it to make the points of access easier for low income Canadians to access their benefits. But it does suggest that there's no magic wand here.
I wouldn't walk into that discussion thinking I'm gonna solve that in a day.
Tegan: Don't get me wrong. All of this isn't a criticism of the Market Basket Measure. It's a tool designed for a specific purpose that doesn't always work when it's taken out of context and applied to a new purpose for which it wasn't designed. The market basket measure is a great tool to measure one key aspect of poverty, but isn't a perfect measurement of all of the areas of need in the country. It's just one indicator, and there are many more.
What are some of the issues that can arise when we only use one measurement to capture something as complex as poverty?
Kevin: The different measures capture different elements of things. So there's nothing measure called the low income measure, which looks at how family is doing relative to the typical family, takes the median family income in Canada and compare draws a line based on that. So that one measure sometimes has kind of a different picture because it bases on kind of compares the low end to the middle. And so what's interesting about that measure is in the 1990s, the median income. The typical family in Canada, their income was actually falling. What that meant was, uh, the low income measure, which was keyed to that median income, was actually getting lower and lower every year, so it was easier to be out of poverty. So we saw poverty measures, poverty outcomes falling based on that one because it was getting easier to pass it. But that doesn't sound like a good thing. If everyone's income is falling. But the low income is falling slightly different speed than the median income. It just doesn't sound good. That's one of the reasons that the Market Basket measure provides a different picture. What's interesting is it doesn't compare like the low to the middle. So that depending on how the middle is doing that change is poverty. It just says, look, what do you need in Canada in 2022 to have an adequate life and so they uh, people StatCan and all of the round tables and discussions have come up with the basket and that doesn't depend on what the median income is doing. It doesn't bounce around it that way. So it's a bit more stable in that way and is arguable. I saw I've measure of deprivation. It's not to say that the low income measure doesn't have its uses just to give an example of that. That is something that's used in international poverty comparisons because whatever is in the Canadian basket for the Market Basket measure makes sense for Canada but might not make sense for Italy. Might not make sense for Japan. They're just gonna have different basket of goods, different cultures, different economies. So international comparisons tend to stick to something like the low income measure, which compares the low to the middle, because that's something you can implement more easily across countries.
Tegan: There's a reason why there are 12 different indicators on the StatCan's poverty hub. One indicator just can't tell the full story on its own.
I feel like I got a little bit of a crash course as a junior policy analyst or something doing this episode.
Kevin: There you go. This is the kind of work a lot of folks in the ESDC and Department of Finance do all the time and trying to design these income programs and it it's a great challenge. It's also very rewarding because when you think about what you're doing, you're trying to help out families who are really in need as best you can. I was using, making great use of Statistics Canada surveys and products to do the best job as we're trying to help design those things. Me on the outside as a researcher and those in government on the inside. But yeah, I think it's a fascinating thing because it's so difficult. It's a real challenge. If it were easy, it'd be done by noon and it would be great. But it's those challenges and trying to find ways of finding wins is a great challenge and one that I have enjoyed working on.
Tegan: As you've said before, if it were easy, somebody would have done it already.
Kevin: Yeah. Yeah, I think that's right.
Tegan: It's difficult to eradicate poverty, but that doesn't mean it's not worth doing. Canada has a poverty reduction strategy in line with the UN's sustainable development goals.
How would eradicating poverty change outcomes for an individual? Well, for somebody who's experiencing poverty, how would that change their lives?
Kevin: So in two ways. So, I would phrase it in terms of having more income for people who are in low income I think it would change in a couple of ways. One is through just the ability to purchase more things that might help out in sustaining there well-being, so this would be better food, better living circumstance, adequate clothing, things like that. But the other more subtle but perhaps also more important, channel is through thinking of the stress that a family undergoes when they don't have enough income when they don't feel like they their kids can keep up with their neighbors in terms of how they're able to participate in things that school, that kind of stress of having a tight budget is has been shown by research to be quite important to one's own well-being and also to the long run impacts of living in poverty. What people remember if they grew up in poverty, is perhaps some hungry nights, but more often than that, it was the uuh, you know, pain in the stress of having a tight budget, whether that led to. I you know, bad behavior in the household or just a shame and embarrassment at school, those memories and the real impact of that is in certain ways even stronger than going at night without a big dinner.
Tegan: In a previous episode, "Unravelling," from season one, our guest, Dr. Kelley, spoke about the impact of stress on kids. Check it out to learn more.
What does it cost not to do our best to eradicate poverty?
Evelyn: Ohh, I think the cost of poverty are immense in this country. Umm, I think, if you start looking at, there's not a single social problem in Canada that's not made worse by poverty. If you look at the healthcare system, for example in 2010. There was a study done on hospitalizations, for example. And I'm looking specifically at what are called ambulatory care sensitive conditions. Now, these are hospitalizations that occur because people didn't receive appropriate primary care. Umm, looking specifically at those kinds of hospitalizations, the authors found that 30 to 40% of hospitalizations are driven by low socioeconomic status. So driven by poverty. If you look at education, a lot of educational funding is required to help children stay current when they change schools. Why don't you children who change schools many times over the course of a year? Why do children change schools? One of the reasons is that parents can't pay the rent and they move. And so you get that kind of mobility among families that and that makes it harder for kids to stay current. And it makes it harder for the educational system to pay for kids. If you look at something like incarceration, 80% of women who are incarcerated. Are incarcerated for poverty related crimes. 80% I mean the cost of incarceration are immense in this country. So we're paying for poverty, we're paying for it in, in, in terms of every social program you can think of. It's not just the money we put out in terms of provincial social assistance or other kinds of programs. It's every single social program that we've gotten place to assist people.
Tegan: And this might be it. Just a question that's not answerable. But how do we measure the worth of breaking the cycle of poverty for a family?
Evelyn: I think that's your rhetorical question to end with. Well, I yeah, I don't have an answer for that. I don't have an answer for that because I think ultimately it's a moral question. It's an ethical question. Umm, you know, it's in the in a sense it, In a sense, I think it undermines the importance of the issue. If I say, well, I benefit, I benefit. If I'm not living. You know next door to people who need help and don't receive it. We all benefit, I think, but we've benefit in in very practical and monetary terms. But we especially benefit, I think, in broader social terms, in terms of the kind of cohesion, the kind of society we want to live in.
Tegan: If someone would like to learn more about the Market Basket measure and how Statcan measure poverty, where can they go?
Burton: So the dimensions of poverty hub at Statistics Canada is a great resource for the latest information on the Market Basket measure and the work that we're doing on creating the Market Basket measure thresholds for the territories that different indicators of poverty that are identified in the opportunity for all Canada's first Poverty reduction document. So I would start there.
Tegan: You've been listening to Eh Sayers. Thank you to our guests, Burton Gustajtis, Evelyn Forget, and Kevin Milligan for sharing their expertise.
You can subscribe to this show wherever you get your podcasts. There, you can also find the French version of our show, called Hé-coutez bien. If you liked this show, please rate, review, and subscribe. Thanks for listening!
Sources
"Dimensions of Poverty Hub." 2018. Statistics Canada. Statistics Canada. December 4, 2018. Dimensions of Poverty Hub
Forget, Evelyn L. 2011. "The Town with No Poverty: The Health Effects of a Canadian Guaranteed Annual Income Field Experiment." Canadian Public Policy 37 (3): 283–305. The Town with No Poverty: The Health Effects of a Canadian Guaranteed Annual Income Field Experiment.
"The Daily — Canadian Income Survey, 2020." 2022. Statistics Canada. March 23, 2022. Canadian Income Survey, 2020.
| NAPCS-CANADA | Month | |||
|---|---|---|---|---|
| 202204 | 202205 | 202206 | 202207 | |
| Total commodities, retail trade commissions and miscellaneous services | 0.67 | 0.63 | 0.61 | 0.72 |
| Retail Services (except commissions) [561] | 0.67 | 0.63 | 0.61 | 0.71 |
| Food at retail [56111] | 0.94 | 0.56 | 0.52 | 1.85 |
| Soft drinks and alcoholic beverages, at retail [56112] | 0.63 | 0.59 | 0.61 | 0.71 |
| Cannabis products, at retail [56113] | 0.00 | 0.00 | 0.00 | 0.00 |
| Clothing at retail [56121] | 1.05 | 1.00 | 0.93 | 0.84 |
| Footwear at retail [56122] | 1.76 | 1.51 | 1.22 | 1.55 |
| Jewellery and watches, luggage and briefcases, at retail [56123] | 7.38 | 5.44 | 5.89 | 5.00 |
| Home furniture, furnishings, housewares, appliances and electronics, at retail [56131] | 1.14 | 1.31 | 1.05 | 1.02 |
| Sporting and leisure products (except publications, audio and video recordings, and game software), at retail [56141] | 2.09 | 1.60 | 1.93 | 1.80 |
| Publications at retail [56142] | 5.82 | 5.62 | 6.05 | 5.64 |
| Audio and video recordings, and game software, at retail [56143] | 0.62 | 0.31 | 1.17 | 1.01 |
| Motor vehicles at retail [56151] | 2.33 | 2.21 | 2.14 | 2.32 |
| Recreational vehicles at retail [56152] | 5.72 | 6.99 | 2.88 | 3.73 |
| Motor vehicle parts, accessories and supplies, at retail [56153] | 1.74 | 1.83 | 1.84 | 1.85 |
| Automotive and household fuels, at retail [56161] | 1.68 | 1.86 | 1.61 | 1.65 |
| Home health products at retail [56171] | 2.39 | 2.54 | 2.58 | 2.53 |
| Infant care, personal and beauty products, at retail [56172] | 2.07 | 1.97 | 2.25 | 1.99 |
| Hardware, tools, renovation and lawn and garden products, at retail [56181] | 2.81 | 1.60 | 2.41 | 2.09 |
| Miscellaneous products at retail [56191] | 3.02 | 3.12 | 2.89 | 2.40 |
| Total retail trade commissions and miscellaneous services Footnote 1 | 1.66 | 1.84 | 1.88 | 1.94 |
Footnotes
|
||||
In October 2022, questions measuring the Labour Market Indicators were added to the Labour Force Survey as a supplement.
Question wording within the collection application is controlled dynamically based on responses provided throughout the survey.
ENTRY_Q01 / EQ1 – From the following list, please select the household member that will be completing this questionnaire on behalf of the entire household.
WFH_Q01 / EQ2 – At the present time, in which of the following locations (do/does) (Respondent’s name/this person/you) usually work as part of (his/her/their/your) main job or business?
WFH_Q02 / EQ3 – Last week, what proportion of (his/her/their/your) work hours did (Respondent name/this person/you) work at home as part of (his/her/their/your) main job or business?
INF_Q01 / EQ4 – Over the last month, that is since September 15 to today, how many hours of voluntary ( /paid) overtime or ( /paid) extra hours did (respondent’s name/this person/you) decide to work at any (business/businesses/job/jobs) in response to the recent increase in the cost of living?
INF_Q02 / EQ5 – When did (respondent’s name/this person/you) last receive a raise in (his/her/their/your) main job?
CHS_Q01 / EQ6 – Over the last month, that is since September 15 to today, how difficult or easy was it for your household to meet its financial needs in terms of transportation, housing, food, clothing and other necessary expenses?
CHS_Q02 / EQ7 – Today, could your household cover an unexpected expense of $500 from your household's resources?
| NAPCS-CANADA | Quarter | |
|---|---|---|
| 2022Q1 | 2022Q2 | |
| Total commodities, retail trade commissions and miscellaneous services | 1.17 | 0.93 |
| Retail Services (except commissions) [561] | 1.20 | 0.95 |
| Food at retail [56111] | 0.97 | 1.66 |
| Soft drinks and alcoholic beverages, at retail [56112] | 0.49 | 0.68 |
| Cannabis products, at retail [56113] | 0.00 | 0.00 |
| Clothing at retail [56121] | 1.25 | 2.96 |
| Footwear at retail [56122] | 1.50 | 2.61 |
| Jewellery and watches, luggage and briefcases, at retail [56123] | 6.58 | 6.12 |
| Home furniture, furnishings, housewares, appliances and electronics, at retail [56131] | 1.45 | 1.81 |
| Sporting and leisure products (except publications, audio and video recordings, and game software), at retail [56141] | 1.96 | 3.25 |
| Publications at retail [56142] | 5.80 | 7.06 |
| Audio and video recordings, and game software, at retail [56143] | 0.50 | 1.04 |
| Motor vehicles at retail [56151] | 1.86 | 1.78 |
| Recreational vehicles at retail [56152] | 3.65 | 3.02 |
| Motor vehicle parts, accessories and supplies, at retail [56153] | 1.62 | 1.63 |
| Automotive and household fuels, at retail [56161] | 1.90 | 1.60 |
| Home health products at retail [56171] | 2.10 | 2.59 |
| Infant care, personal and beauty products, at retail [56172] | 2.20 | 3.55 |
| Hardware, tools, renovation and lawn and garden products, at retail [56181] | 2.14 | 2.08 |
| Miscellaneous products at retail [56191] | 2.00 | 3.08 |
| Total retail trade commissions and miscellaneous services Footnotes 1 | 1.76 | 1.57 |
Footnotes
|
||
In October 2022, the following questions measuring the Labour Market and Socioeconomic Indicators were added to the Labour Force Survey as a supplement.
The purpose of this survey is to identify changing dynamics within the Canadian labour market, and measure important socioeconomic indicators by gathering data on topics such as type of employment, quality of employment, support payments and unmet health care needs.
Question wording within the collection application is controlled dynamically based on responses provided throughout the survey.
ENTRY_Q01 / EQ 1 - From the following list, please select the household member that will be completing this questionnaire on behalf of the entire household.
Employee Block
LMI_Q01 / EQ 2 - Is (respondent name's/this person's/your) main job permanent?
LMI_Q02 / EQ 3 - In what way is (respondent name's/this person's/your) main job not permanent?
LMI_Q03 / EQ 4 - In (his/her/their/your) main job, (are/is) (respondent name/this person/you) paid by a private employment or placement agency that is different from the company (he/she/this person/you) work(s) for?
LMI_Q04 / EQ 5 - What is the total duration of (respondent name's/this person's/your) contract or agreement in (his/her/their/your) main job?
LMI_Q05 / EQ 6 - In (respondent name's/this person's/your) main job, (is/are) (he/she/they/you) guaranteed a minimum number of work hours per pay period?
LMI_Q06 / EQ 7 - What would you say best describes (respondent name's/this person's/your) current situation in (his/her/their/your) main job?
Self-employed Block
LMI_Q07 / EQ 8 – What is the main reason why (respondent name/this person/you) (are/is) self-employed in (his/her/their/your) (main/other) job?
LMI_Q08 / EQ 9 – (Does/do) (respondent name/this person/you) have any partners or co-owners in (his/her/their/your) (main/side) business?
LMI_Q09 / EQ 10 – (Does/Do) (respondent name/this person/you) or (his/her/their/your) (partners/company/partners or company) own or lease a building or space dedicated to (his/her/their/your) (main/side) business?
LMI_Q10 / EQ 11 - In (respondent name's/this person's/your) (main/side) business, is/are (he/she/they/you) required to belong to a professional association or regulatory college to do (his/her/their/your) job?
LMI_Q11 / EQ 12 - Does (respondent name's/this person's/your) (main/side) business operate…?
LMI_Q12 / EQ 13 - What is the current mix of clients in (respondent name's/this person's/your) main business?
LMI_Q13 / EQ 14 - Would (respondent name/this person/you) be able to continue operating (his/her/their/your) main business for the next five years based on returning or existing clients alone?
LMI_Q14 / EQ 15 – To what extent do you agree or disagree with the following statement?
In normal times, it is easy for [Respondent's name/this person/you] to find new clients in [his/her/their/your] main business.
LMI_Q15 / EQ 16 – (Does/Do) (respondent name/this person/you) or (his/her/their/your) (partners/company/partners or company) currently have any contracts with businesses, government agencies, or non-profit organizations as part of (his/her/their/your) main business?
LMI_Q16 / EQ 17 - Thinking of (respondent name's/this person's/your) largest contract, what is the total duration of that contract?
LMI_Q17 / EQ 18 – During the last 12 months, did (respondent name/this person/you) have any full days with no clients or work in (his/her/their/your) main business even though (he/she/they/you) wanted to work?
LMI_Q18 / EQ 19 – What would you say is (respondent name's/this person's/your) plan with (his/her/their/your) main business over the next 12 months?
LMI_Q19 / EQ 20 - What is the main reason (respondent name/this person/you) expect(s) to stop working or close (his/her/their/your) main business?
LFI-CHECK1 / EQ 21 - Last week, did (he/she/this person/you) work at a job or business?
LFI-CHECK2 / EQ 22 - Last week, did (he/she/this person/you) have a job or business from which (he/she/this person/you) (were/was) absent?
LFI-CHECK3 / EQ 23 - Did (he/she/this person/you) have more than one job or business last week?
LFI-CHECK4 / EQ 24 - Was this because (he/she/this person/you) changed employers?
LFI-CHECK5 / EQ 25 - (Has/Have) (respondent name/this person/you) ever worked at a job or business?
LFI-CHECK6 / EQ 26 - When did (respondent name/this person/you) last work?
LMI_Q20 / EQ 27 - Excluding (his/her/their/your) main job or business, (has/have) (respondent's name/this person/you) earned any money by freelancing, doing a paid gig, or completing a short-term job or task during the last 12 months?
LMI_Q21 / EQ 28 - Was this freelancing, paid gig, or short-term task or job one of the jobs (respondent's name/this person/you) had last week, or something else entirely?
LMI_Q22 / EQ 29 - You mentioned earlier that (respondent name/this person/you) had a job or a business during the last 12 months.
In addition to the jobs or businesses (he/she/this person/you) had during this period, did (respondent name/this person/you) earn any money by freelancing, doing a paid gig, or completing any other short-term job or task during the last 12 months?
LMI_Q23 / EQ 30 – (You mentioned earlier that (respondent name/this person/you) did not have a job or a business during the last 12 months.)
(Has/Have) (respondent name/this person/you) earned any money by freelancing, doing a paid gig, or completing any other short-term job or task during the last 12 months?
LMI_Q24 / EQ 31 - When was the last time (respondent name/this person/you) freelanced, did a paid gig, or got paid to do a short-term task or job?
SCC1_Q05 / EQ 32 - In the last 12 months, did (respondent's name/you) receive support payments from a former spouse or partner?
SCC1_Q10 / EQ 33 - What is your best estimate of the amount of support payments (he/she/this person/you) received in the last 12 months?
SCC2_Q05 / EQ 34 - In the last 12 months, did (respondent's name/you) make support payments to a former spouse or partner?
SCC2_Q10 / EQ 35 - What is your best estimate of the total amount (he/she/this person/you) paid in support payments in the last 12 months?
SCC3_Q05 / EQ 36 - In the last 12 months, did (respondent's name/you) pay for child care, so that (he/she/they/you) could work at a paid job?
SCC3_Q10 / EQ 37 - What is your best estimate, of the total amount (he/she/this person/you) paid for child care in the last 12 months?
DSQ_Q01 / EQ 38 - (Do/Does) (respondent's name/you) have any difficulty seeing?
DSQ_Q02 / EQ 39 - (Do/Does) (he/she/this person/you) wear glasses or contact lenses to improve (respondent name's/this person's/your) vision?
DSQ_Q03 / EQ 40 - (Which/With (respondent name's/this person's/your) glasses or contact lenses, which) of the following best describes (respondent's name/your) ability to see?
DSQ_Q04 / EQ 41 - How often does this (difficulty seeing/seeing condition) limit (his/her/their/your) daily activities?
DSQ_Q05 / EQ 42 - (Do/Does) (respondent's name/you) have any difficulty hearing?
DSQ_Q06 / EQ 43 - (Do/Does) (he/she/this person/you) use a hearing aid or cochlear implant?
DSQ_Q07 / EQ 44 - (Which/With) (respondent name's/this person's/your) hearing aid or cochlear implant, which) of the following best describes (respondent's name/your) ability to hear?
DSQ_Q08 / EQ 45 - How often does this (difficulty hearing/hearing condition) limit (his/her/their/your) daily activities?
DSQ_Q09 / EQ 46 - (Do/Does) (respondent's name/you) have any difficulty walking, using stairs, using (his/her/their/your) hands or fingers or doing other physical activities?
DSQ_Q10 / EQ 47 - How much difficulty (do/does) (he/she/this person/you) have walking on a flat surface for 15 minutes without resting?
DSQ_Q11 / EQ 48 - How much difficulty (do/does) (he/she/this person/you) have walking up or down a flight of stairs, about 12 steps without resting?
DSQ_Q12 / EQ 49 - How often (does this difficulty walking/does this difficulty using stairs/do these difficulties) limit (his/her/their/your) daily activities?
DSQ_Q13 / EQ 50 - How much difficulty (do/does) (respondent's name/you) have bending down and picking up an object from the floor?
DSQ_Q14 / EQ 51 - How much difficulty (do/does) (he/she/this person/you) have reaching in any direction, for example, above (his/her/their/your) head?
DSQ_Q15 / EQ 52 - How often (does this difficulty bending down and picking up an object/does this difficulty reaching/do these difficulties) limit (his/her/their/your) daily activities?
DSQ_Q16 / EQ 53 - How much difficulty (do/does) (respondent's name/you) have using (his/her/their/your) fingers to grasp small objects like a pencil or scissors?
DSQ_Q17 / EQ 54 - How often does this difficulty using (his/her/their/your) fingers limit (his/her/their/your) daily activities?
DSQ_Q18 / EQ 55 - (Do/Does) (respondent's name/you) have pain that is always present?
DSQ_Q19 / EQ 56 - (Do/Does) (he/she/this person/you) ( /also) have periods of pain that reoccur from time to time?
DSQ_Q20 / EQ 57 - How often does this pain limit (his/her/their/your) daily activities?
DSQ_Q21 / EQ 58 - When (respondent's name/you) (are/is) experiencing this pain, how much difficulty (do/does) (he/she/they/you) have with (his/her/their/your) daily activities?
DSQ_Q22 / EQ 59 - (Do/Does) (respondent's name/you) have any difficulty learning, remembering or concentrating?
DSQ_Q23 / EQ 60 - Do you think (respondent's name/you) (has/have) a condition that makes it difficult in general for (him/her/them/you) to learn? This may include learning disabilities such as dyslexia, hyperactivity, attention problems, etc.
DSQ_Q24 / EQ 61 - Has a teacher, doctor or other health care professional ever said that (respondent's name/you) had a learning disability?
DSQ_Q25 / EQ 62 - How often are (his/her/their/your) daily activities limited by this condition?
DSQ_Q26 / EQ 63 - How much difficulty (do/does) (respondent's name/you) have with (his/her/their/your) daily activities because of this condition?
DSQ_Q27 / EQ 64 - Has a doctor, psychologist or other health care professional ever said that (respondent's name/you) had a developmental disability or disorder? This may include Down syndrome, autism, Asperger syndrome, mental impairment due to lack of oxygen at birth, etc.
DSQ_Q28 / EQ 65 - How often are (respondent's name/your) daily activities limited by this condition?
DSQ_Q29 / EQ 66 - How much difficulty (do/does) (respondent's name/you) have with (his/her/their/your) daily activities because of this condition?
DSQ_Q30 / EQ 67 - (Do/Does) (he/she/this person/you) have any ongoing memory problems or periods of confusion?
DSQ_Q31 / EQ 68 - How often are (his/her/their/your) daily activities limited by this problem?
DSQ_Q32 / EQ 69 - How much difficulty (do/does) (respondent's name/you) have with (his/her/their/your) daily activities because of this problem?
DSQ_Q33 / EQ 70 - (Do/Does) (respondent's name/you) have any emotional, psychological or mental health conditions?
DSQ_Q34 / EQ 71 - How often are (his/her/their/your) daily activities limited by this condition?
DSQ_Q35 / EQ 72 - When (respondent's name/you) (are/is) experiencing this condition, how much difficulty (do/does) (he/she/they/you) have with (his/her/their/your) daily activities?
DSQ_Q36 / EQ 73 - (Do/Does) (respondent's name/you) have any other health problem or long-term condition that has lasted or is expected to last for six months or more?
DSQ_Q37 / EQ 74 - How often does this health problem or long-term condition limit (his/her/their/your) daily activities?
DSQ_Q38 / EQ 75 - (Do/Does) (respondent's name/you) have pain that is always present?
DSQ_Q39 / EQ 76 - (Do/Does) (he/she/this person/you) ( /also) have periods of pain that reoccur from time to time?
DSQ_Q40 / EQ 77 - How often does this pain limit (his/her/their/your) daily activities?
DSQ_Q41 / EQ 78 - When (respondent's name/you) (are/is) experiencing this pain, how much difficulty (do/does) (he/she/they/you) have with (his/her/their/your) daily activities?
UNC_Q005 / EQ 79 - During the past 12 months, was there ever a time when (respondent's name/you) felt that (he/she/they/you) needed health care, other than homecare services, but (he/she/they/you) did not receive it?
UNC_Q010 / EQ 80 - Thinking of the most recent time (respondent's name/you) felt this way, why didn't (he/she/they/you) get care?
UNC_Q015 / EQ 81 - Again, thinking of the most recent time, what was the type of care that was needed?
UNC_Q020 / EQ 82 - Did (he/she/this person/you) actively try to obtain the health care that was needed?
UNC_Q025 / EQ 83 - Where did (he/she/this person/you) try to get the service (he/she/they/you) (was/were) seeking?
1. Which of the following categories best describes this business or organization?
2. In what year was this business or organization first established?
Please provide the year this business or organization first began operations.
Year business or organization was first established:
OR
Don't know
3. Over the last 12 months, did this business or organization conduct any of the following international activities?
Select all that apply.
4. Over the next three months, how are each of the following expected to change for this business or organization?
Exclude seasonal factors or conditions.
5. Over the next three months, which of the following are expected to be obstacles for this business or organization?
Select all that apply.
Flow condition: If "Recruiting skilled employees" or "Retaining skilled employees" is selected in Q5, go to Q6. Otherwise, go to Q7.
6. Compared with 12 months ago, how would this business or organization describe its challenges with recruiting and retaining staff?
Flow condition: If "Difficulty acquiring inputs, products or supplies from within Canada", "Difficulty acquiring inputs, products or supplies from abroad", or "Maintaining inventory levels" is selected in Q5, go to Q7. Otherwise, go to Q12.
7. How long does this business or organization expect the following to continue to be an obstacle?
8. Over the last three months, how have supply chain challenges experienced by this business or organization changed?
Supply chain challenges include difficulty acquiring inputs, products or supplies from within Canada or abroad and difficulty maintaining inventory levels.
Exclude seasonal factors or conditions.
9. Over the next three months, how does this business or organization expect supply chain challenges to change?
Supply chain challenges include difficulty acquiring inputs, products or supplies from within Canada or abroad and difficulty maintaining inventory levels.
Exclude seasonal factors or conditions.
10. Over the next 12 months, does this business or organization plan to make any of the following adjustments to its supply chain?
Select all that apply.
Display condition: If "Maintaining inventory levels" is selected in Q5, go to Q11. Otherwise, go to Q12.
11. Over the next three months, in response to an expected difficulty maintaining inventory levels, which of the following does this business or organization plan to do?
Select all that apply.
Flow condition: If the business or organization is a private sector business or non-profit organization, go to Q12. Otherwise, go to Q13.
Display condition: If the business or organization is a non-profit organization, do not display "Transfer the business" or "Sell the business".
12. Over the next 12 months, does this business or organization plan to do any of the following?
Select all that apply.
13. Over the next 12 months, how likely is this business or organization to pass on any increases in its costs to customers?
e.g., costs related to increases in wages, inputs, products, supplies, taxes, rents, and carbon prices.
14. What percentage of employees does this business or organization expect to voluntarily retire over the next 12 months?
Exclude layoffs.
Provide your best estimate rounded to the nearest percentage.
Percentage of employees expected to retire over the next 12 months:
OR
Don't know
Flow condition: If the percentage of employees expected to retire over the next 12 months is greater than 0% in Q14, go to Q15. Otherwise, go to Q16.
15. Does this business or organization have plans in place to address expected retirements?
e.g., hiring staff to fill vacancies due to retirements, training staff to take over responsibilities of retiring staff
16. Over the next 12 months, does this business or organization expect the average wages paid to change?
17. To what extent does this business or organization consider inflation when setting wages and salaries?
18. Does this business or organization currently do or plan to do any of the following over the next 12 months?
Select all that apply.
19. Does this business or organization have, or intend to recruit, volunteers?
Display condition: If "Yes" is selected in Q19, go to Q20. Otherwise, go to Q21.
20. Volunteer recruitment and retention challenges have had which of the following impacts or expected impacts on this business or organization?
Select all that apply.
21. Over the next 12 months, does this business or organization plan to adopt or incorporate any of the following technologies?
Exclude technologies already adopted.
Select all that apply.
Flow condition: If at least one technology or "Other" is selected in Q21, go to Q22. Otherwise, go to Q23.
22. Using a scale from 1 to 5, where 1 means "not at all challenging" and 5 means "extremely challenging", how challenging are the following for this business or organization when adopting or incorporating technologies?
23. Over the last 12 months, did this business or organization apply for a new line of credit, a new term loan, a new non-residential mortgage, or refinancing of an existing non-residential mortgage?
Include commercial mortgages.
Exclude residential mortgages.
24. Does this business or organization have the cash or liquid assets required to operate for the next three months?
25. Over the next three months, does this business or organization plan to apply for a new line of credit, a new term loan, a new non-residential mortgage, or refinancing of an existing non-residential mortgage?
Include commercial mortgages.
Exclude residential mortgages.
26. Which of the following best describes the current debt level of this business or organization?
27. Over the next 12 months, to what extent does this business or organization foresee challenges in repaying funding received from repayable government support programs put in place because of the COVID-19 pandemic?
Examples of repayable government support programs include the Canada Emergency Business Account (CEBA) or the Indigenous Business Initiative (sometimes referred to as the Emergency Loan Program (ELP), issued through an Aboriginal Financial Institutions (AFI) or Métis Capital Corporations (MCCs)).
28. Over the next three months, what percentage of the employees of this business or organization is anticipated to do each of the following?
Exclude staff that are primarily engaged in providing driving or delivery services or staff that primarily work at client premises.
Provide your best estimate rounded to the nearest percentage.
If the percentages are unknown, leave the question blank.
29. Over the next 12 months, what is the future outlook for this business or organization?
Flow condition: If the business or organization is a private sector business, go to Q30. Otherwise, go to "Contact person".
(i) The groups identified within the following questions are included in order to gain a better understanding of businesses owned by members of various communities across Canada.
30. What percentage of this business or organization is owned by women?
Provide your best estimate rounded to the nearest percentage.
Percentage:
OR
Don't know
31. What percentage of this business or organization is owned by First Nations, Métis or Inuit peoples?
Provide your best estimate rounded to the nearest percentage.
Percentage:
OR
Don't know
32. What percentage of this business or organization is owned by immigrants to Canada?
Provide your best estimate rounded to the nearest percentage.
Percentage:
OR
Don't know
33. What percentage of this business or organization is owned by persons with a disability?
Include visible and non-visible disabilities.
Provide your best estimate rounded to the nearest percentage.
Percentage:
OR
Don't know
34. What percentage of this business or organization is owned by LGBTQ2 individuals?
The term LGBTQ2 refers to persons who identify as lesbian, gay, bisexual, transgender, queer and/or two-spirited.
Provide your best estimate rounded to the nearest percentage.
Percentage:
OR
Don't know
35. What percentage of this business or organization is owned by members of visible minorities?
A member of a visible minority in Canada may be defined as someone (other than an Indigenous person) who is non-white in colour or race, regardless of place of birth.
Provide your best estimate rounded to the nearest percentage.
Percentage:
OR
Don't know
Flow condition: If more than 50% of this business or organization is owned by members of visible minorities, go to Q36. Otherwise, go to "Contact person".
36. It was indicated that at least 51% of this business or organization is owned by members of visible minorities. Please select the categories that describe the owner or owners.
Select all that apply.
By: Mihir Gajjar, Statistics Canada
Contributors: Reginald Maltais, Allie Maclsaac, Claudia Mokbel and Jeremy Solomon, Statistics Canada
MLflow is an open source platform that manages the machine learning lifecycle, including experimentation, reproducibility, deployment and a central model registry. MLflow offers four components:
This article focuses on MLflow Tracking. The MLflow website has details on the remaining three components.
MLflow Tracking provides a solution that can be scaled from your local machine to the entire enterprise. This allows data scientists to get started on their local machine while organizations can implement a solution that ensures long term maintainability and transparency in a central repository.
MLflow Tracking provides consistent and transparent tracking by:
This article focuses on using MLflow with Python. The MLflow QuickStart document has examples of its use with R for a local installation on a single machine. Organizations wishing to deploy MLflow across teams could also refer to the QuickStart document.
This article will explore an example of using MLflow with Python; however, to get the best understanding of how MLFlow works, it's useful to go through each step on your machine.
MLflow can be installed as a standard Python package by typing the following command in a terminal window:
$ pip install mlflow
After the command has finished executing, you can type mlflow in your terminal and explore the available options. For example, you can try: mlflow –version to see the version installed
It's recommended to have a centralized MLflow server for an individual, team or organization so that runs for different projects can be logged in one central place, segregated by experiments (different experiments for different projects). This will be covered in more detail later in this article. To quickly get started with the tool, you can skip the server launch and still log the runs. By doing this, the runs are stored in a directory called "MLruns" located in the same directory as the code. You can later open MLflow UI in the same path and visualize the logged runs.
The runs can be logged to an MLflow server running locally or remotely by setting the appropriate tracking URI (uniform resource identifier). Setting the appropriate logging location is explained later.
If, however, you prefer to start the server right away, you can do so by issuing the following command:
$ mlflow server
The terminal will display information similar to what is below, which shows the server is listening at localhost port 5000This address is useful for accessing the MLflow user interface (UI). Feel free to explore the subtle difference between MLflow UI and MLflow server in the MLflow Tracking documentation.
[2021-07-09 16:17:11 –0400] [58373] [INFO] Starting gunicorn 20.1.0
[2021-07-09 16:17:11 –0400] [58373] [INFO] Listening at: http://127.0.0.1:5000 (58373)
[2021-07-09 16:17:11 –0400] [58373] [INFO] Using worker: sync
[2021-07-09 16:17:11 –0400] [58374] [INFO] Booting worker with pid: 58374
[2021-07-09 16:17:11 –0400] [58375] [INFO] Booting worker with pid: 58375
[2021-07-09 16:17:11 –0400] [58376] [INFO] Booting worker with pid: 58376
[2021-07-09 16:17:11 –0400] [58377] [INFO] Booting worker with pid: 58377
There are two main concepts in MLflow tracking: experiments and runs. The data logged during an experiment is recorded as a run in MLflow. The runs can be organized into experiments, which groups together runs for a specific task. One can visualize, search, compare, and download run artifacts and metadata for the runs logged in an MLflow experiment.
Data in an experiment can be logged as a run in MLflow using MLflow Python, R, Java packages, or through the REST API (application programming interface).
This article will demonstrate modeling for one of the "Getting started with NLP (natural language processing)" competitions on Kaggle called "Natural Language Processing with Disaster Tweets." A Jupyter notebook and the MLflow Python API will be used for logging data to MLflow. The focus will be on demonstrating how to log data to MLflow during modeling, rather than getting the best modeling results.
First, let's start with the usual modeling process, which includes imports, reading the data, text pre-processing, tf-idf (term frequency-inverse document frequency) features and support vector machine (SVM) model. At the end, there will be a section called "MLflow logging."
Note: The NLP pipeline is kept as simple as possible so that the focus is on MLflow logging. Some of the usual steps, like exploratory data analysis, are not relevant for this purpose and will be left out. The preferred way of logging data to MLflow is by leaving a chunk of code at the end to log. You can also configure MLflow at the beginning of the code and log data throughout the code, when the data or variable is available to log. An advantage to logging all the data together at the end using a single cell is that the entire pipeline would finish successfully, and the run will log the data (given the code for MLflow logging has no bugs). If the data are logged throughout the code and the code execution stops for any reason, the data logging will be incomplete. However, if there's a scenario where a code has more than one code chunk, which takes a significant amount of time to execute, then logging throughout the code, in multiple locations, may actually be beneficial.
Start by importing all the required libraries for the example:
# To create unique run name.
import time
# To load data in pandas dataframe.
import pandas as pd
# NLP libraries
# To perform lemmatization
from nltk import WordNetLemmatizer
# To split text into words
from nltk. tokenize import word_tokenize
# To remove the stopwords
from nltk.corpus import stopwords
# Scikit-learn libraries
# To use the SVC model
from sklearn.svm import SVC
# To evaluate model performance
from sklearn.model_selection import cross_validate, StratifiedkFold
# To perform Tf-idf vectorization
from sklearn.feature_extraction.text import TfidfVectorizer
# To get the performance metrics
from sklearn.metrics import f1_score, make_scorer
# For logging and tracking experiments
import mlflow
MLflow tracks multiple runs of an experiment through a run name parameter. The run name can be set to any value, but should be unique so you can identify it amongst different runs later. Below, a timestamp is used to guarantee a unique name.
run_name = str(int(time.time()))
print('Run name: ', run_name)
Gives:
Run name: 1625604741Load the training and test data from the CSV files provided by the example.
# Kaggle competition data download link: https://www.kaggle.com/c/nlp-getting-started/data
train_data = pd.read_csv("./data/train.csv")
test_data = pd.read_csv("./data/test.csv")
By executing the following piece of code in a cell:
train_data
A sample of the training data that was just loaded can be seen in Figure 1.
Figure 1: A preview of the training data that was loaded.
The top and bottom five entries of the CSV file. It contains the columns: id, keyword, location, text and target. The text column contains the tweet itself and the target column, the class.
| id | keyboard | location | text | target | |
|---|---|---|---|---|---|
| 0 | 1 | NaN | Nan | Our Deeds are the Reason of this #earthquake M… | 1 |
| 1 | 4 | NaN | Nan | Forest fire near La Ronge Sask. Canada | 1 |
| 2 | 5 | NaN | Nan | All residents asked to 'shelter in place' are… | 1 |
| 3 | 6 | NaN | Nan | 13,000 people receive #wildfires evacuation or… | 1 |
| 4 | 7 | NaN | Nan | Just got sent this photo from Ruby #Alaska as… | 1 |
| ... | ... | ... | ... | ... | ... |
| 7608 | 10869 | NaN | Nan | Two giant cranes holding a bridge collapse int… | 1 |
| 7609 | 10870 | NaN | Nan | @aria_ahrary @TheTawniest The out of control w… | 1 |
| 7610 | 10871 | NaN | Nan | M1.94 [01:04 UTC] ?5km S of Volcano Hawaii. Htt… | 1 |
| 7611 | 10872 | NaN | Nan | Police investigating after an e-bike collided… | 1 |
| 7612 | 10873 | NaN | Nan | The Latest: More Homes Razed by Northern Calif… | 1 |
7613 rows x 5 columns
The training data are about 70% of the total data.
print('The length of the training data is %d' % len(train_data))
print('The length of the test data is %d' % len(test_data))
Output:
The length of the training data is 7613
The length of the test data is 3263
Depending on the task at hand, different types of preprocessing steps might be required to make the machine learning model learn better features. Preprocessing can normalize the input, remove some of the common words if required so that the model does not learn them as features, make logical and meaningful changes that can lead to the model performing and generalizing better. The following demonstrates how performing some pre-processing steps can help the model grab the right features when learning:
def clean_text(text):
# split into words
tokens = word_tokenize(text)
# remove all tokens that are not alphanumeric. Can also use .isalpha() here if do not want to keep numbers.
words = [word for word in tokens if word.isalnum()]
# remove stopwords
stop_words = stopwords.words('english')
words = [word for word in words if word not in stop_words]
# performing lemmatization
wordnet_lemmatizer = WordNetLemmatizer()
words = [wordnet_lemmatizer.lemmatize(word) for word in words]
# Converting list of words to string
words = ' '.join(words)
return words
train_data['cleaned_text'] = train_data['text'].apply(clean_text)
Comparing the original text to the cleaned text, non-words have been removed:
train_data['text'].iloc[100]
'.@NorwayMFA #Bahrain police had previously died in a road accident they were not killed by explosion https://t.co/gFJfgTodad'
train_data['cleaned_text'].iloc[100]rain_data['text'].iloc[100]
'NorwayMFA Bahrain police previously died road accident killed explosion http'
Reading the text above, one can say that, yes, it does contain information about a disaster and hence should be classified as one. To confirm this with the data, print out the label present in the CSV file for this tweet:
train_data['target'].iloc[100]
Output:
1
Next, we are converting a collection of raw documents to a matrix of TF-IDF features to feed into the model. For more information about tf-idf, please refer to tf–idf - Wikipedia and scikit-learn sklearn.feature_extraction.text documentation.
ngram_range=(1,1)
max_features=100
norm='l2'
tfidf_vectorizer = TfidfVectorizer(ngram_range=ngram_range, max_features=max_features, norm=norm)
train_data_tfidf = tfidf_vectorizer.fit_transform(train_data['cleaned_text'])
train_data_tfidf
Output:
<7613x100 sparse matrix of type '<class 'numpy.float64'>'
with 15838 stored elements in Compressed Sparse Row format>
tfidf_vectorizer.get_feature_names()[:10]
Output:
['accident',
'amp',
'and',
'as',
'attack',
'back',
'best',
'body',
'bomb',
'building']
The next step to perform the modeling is to fit a model and evaluate the performance.
Stratified K-Folds cross-validator is used to evaluate the model. See scikit learn sklearn.model_selection for more information.
strat_k_fold = StratifiedKFold(n_splits=5, shuffle=True, random_state=42)
Making a scorer function using the f1-score metric to pass it as a parameter in the SVC model.
scoring_function_f1 = make_scorer(f1_score, pos_label=1, average='binary')
Now comes an important step of fitting the model to the data. This example uses the SVC classifier. See scikit learn sklearn.svm.svc for more information.
C = 1.0
kernel='poly'
max_iter=-1
random_state=42
svc = SVC(C=C, kernel=kernel, max_iter=max_iter, random_state=random_state)
cv_results = cross_validate(estimator=svc, X=train_data_tfidf, y=train_data['target'], scoring=scoring_function_f1, cv=strat_k_fold, n_jobs=-1, return_train_score=True)
cv_results
Output:
{'fit_time': array([0.99043322, 0.99829006, 0.94024873, 0.97373009, 0.96771407]),
'score_time': array([0.13656974, 0.1343472 , 0.13345313, 0.13198996, 0.13271189]),
'test_score': array([0.60486891, 0.65035517, 0.5557656 , 0.5426945 , 0.63071895]),
'train_score': array([0.71281362, 0.76168757, 0.71334394, 0.7291713 , 0.75554698])}
def mean_sd_cv_results(cv_results, metric='F1'):
print(f"{metric} Train CV results: {cv_results['train_score'].mean().round(3)} +- {cv_results['train_score'].std().round(3)}")
print(f"{metric} Val CV results: {cv_results['test_score'].mean().round(3)} +- {cv_results['test_score'].std().round(3)}")
mean_sd_cv_results(cv_results)
F1 Train CV results: 0.735 +- 0.021
F1 Val CV results: 0.597 +- 0.042
Note: The code below is executed as a shell command by adding the exclamation mark: '!' in the beginning of the code in a Jupyter cell.
! Jupyter nbconvert --to html mlflow-example-real-or-not-disaster-tweets-modeling-SVC.ipynb
[NbConvertApp] Converting notebook mlflow-example-real-or-not-disaster-tweets-modeling-SVC.ipynb to html
[NbConvertApp] Writing 610630 bytes to mlflow-example-real-or-not-disaster-tweets-modeling-SVC.htmlFirst, set the server URI. As the server is running locally, set the tracking URI to localhost port 5000. The tracking URI can be set to a remote server as well (see Where Runs are Recorded).
server_uri = 'http://127.0.0.1:5000'
mlflow.set_tracking_uri(server_uri)To organize the runs, an experiment was created and set where the runs will be logged. The "set_experiment" method will create a new run with the given string name and set it as the current experiment where the runs will be logged.
mlflow.set_experiment('nlp_with_disaster_tweets')Finally, start a run and log data to MLflow.
# MLflow logging.
with mlflow.start_run(run_name=run_name) as run:
# Logging tags
# run_name.
mlflow.set_tag(key='Run name', value=run_name)
# Goal.
mlflow.set_tag(key='Goal', value='Check model performance and decide whether we require further pre-processing/hyper-parameter tuning.')
# Modeling exp.
mlflow.set_tag(key='Modeling technique', value='SVC')
# Logging parameters
mlflow.log_param(key='ngram_range', value=ngram_range)
mlflow.log_param(key='max_features', value=max_features)
mlflow.log_param(key='norm', value=norm)
mlflow.log_param(key='C', value=C)
mlflow.log_param(key='kernel', value=kernel)
mlflow.log_param(key='max_iter', value=max_iter)
mlflow.log_param(key='random_state', value=random_state)
# Logging the SVC model.
mlflow.sklearn.log_model(sk_model=svc, artifact_path='svc_model')
# Logging metrics.
# mean F1-score - train.
mlflow.log_metric(key='mean F1-score - train', value=cv_results['train_score'].mean().round(3))
# mean F1-score - val.
mlflow.log_metric(key='mean F1-score - val', value=cv_results['test_score'].mean().round(3))
# std F1-score - train.
mlflow.log_metric(key='std F1-score - train', value=cv_results['train_score'].std().round(3))
# std F1-score - val.
mlflow.log_metric(key='std F1-score - val', value=cv_results['test_score'].std().round(3))
# Logging the notebook.
# Nb.
mlflow.log_artifact(local_path='real-or-not-disaster-tweets-modeling-SVC.ipynb', artifact_path='Notebook')
# Nb in HTML.
mlflow.log_artifact(local_path='real-or-not-disaster-tweets-modeling-SVC.html', artifact_path='Notebook')
In the code above, you begin a run with a run_name and then log the following:
There you have it—you successfully logged data for a run in MLflow! The next step is to visualize the logged data.
If you scroll back to Figure 1, you'll remember that you launched the server and it was listening at localhost port 5000. Open this address in your preferred browser to access the MLflow UI. Once the MLflow UI is visible, you can use the interface to look at the experiment data that was logged. The experiments created appear in the sidebar of the UI, and the logged tags, parameters, model and metrics are shown in the columns.
Figure 2: MLflow UI
Figure 2 shows the MLflow UI. The experiment which was set above i.e. nlp_with_disaster_tweets is opened and the run that you logged earlier along with the details such as run name, parameters and metrics. It also shows the location where the artifacts are stored. You can click on the logged run to explore it in further detail.
Text in image: MLflow Expriements Models
Nlp_with_disaster_tweets (1. Click on this experiment)
Experiment ID: 1, Artifact Location: ./miruns/1 (Location of the logged artifacts)
Notes: None
2. Explore the logged data
| Parameters | Metrics | Tags | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Start Time | Run Name | User | Source | Version | Model | C | Kernel | max_feature | Mean F1-s | Mean F1-s | Std F1-s | Modeling t |
| 2021-07-09 16:27:58 | 16258… | Mihir | ipykerne | - | sklearn | 1.0 | poly | 100 | 0.735 | 0.597 | 0.021 | SVC |
3. The run that we logged using the Python API. Click on the link to open the run
To explore a specific run in greater detail, click on the relevant run in the Start Time column. This will allow you to explore a logged run in detail. The run name is shown and you can add any notes for the run such as logged parameters, metrics, tags and artifacts. The data logged using the Python API for this run are shown here.
The files logged as artifacts can be downloaded, which can be useful if you want to retrieve the code later. Since the code that generates results for every run is saved, you don't need to create multiple copies of the same code and can experiment using a single skeleton notebook by changing the code between runs.
The logged trained model can be loaded in a future experiment using the Python API from the logged run.
Figure 3: Exploring the logged artifacts in a run
Figure 3 explores the logged artifacts. The logged files (notebook and the model) are shown. The description of the model also provides code to load the logged model in Python.
Text in image:
Tags
| Name | Value | Actions |
|---|---|---|
| Goal | Check model performance and decide whether we require further pre-processing hyper-parameter tuning. | Edit – delete icons |
| Modeling technique | SVC | Edit – delete icons |
| Run name | 16525862471 | Edit – delete icons |
Add Tag
Name – Value – Add
Artifacts
Notebook
svc_model
Full Path: ./miruns/1/fcdc8362b2fe74329a4128fa522d80cb/artifacts/svc_model
Size: 0B
MLflow Model
The code snippets below demonstrate how to make predictions using the logged model
Model schema
Input and output schema for your model. Learn more
Name – Type
No Schema.
To demonstrate the run comparison functionality, more modeling experiments were performed and logged to MLflow by changing a few parameters in the same jupyter notebook. Feel free to change some parameters and log more runs to MLflow.
Figure 4 shows the different logged runs. You can filter, keep the columns you want, and compare the parameters or metrics between different runs. To perform a detailed comparison, you can select the runs you want to compare and click on the "Compare" button highlighted in the figure below.
Figure 4: Customizing and comparing different runs using MLflow UI
In MLflow UI, one can customize the columns being shown, filter and search for different runs based on the logged data and can easily compare the different logged runs based on the visible columns. You can also compare different logged runs in greater detail by selecting them and clicking on the "Compare" button.
Text in image:
1. Can filter and keep the columns of interest
Columns: Start Time, User, Run Name, Source, Version, Models, Parameters, Metrics, Tags
2. Can compare different runs
3. Different runs logged. Select the runs you want to compare
Showing 5 matching runs. Compare, Delete, Download, CSV
4. Click Compare
| Parameters | Metrics | Tags | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Start Time | Run Name | User | Source | Version | Model | C | Kernel | max_fe | Mean F1-s | Mean F1-s | Std F1-s | Modeling | Run name | Goal |
| 2021-07-12 14:48:54 | 1626115725 | Mihir | ipykerne | - | sklearn | 1.0 | Poly | 500 | 0.93 | 0.694 | 0.001 | SVC | 16261157 | Check mo… |
| 2021-07-12 14:48:16 | 1626115688 | Mihir | ipykerne | - | sklearn | 1.0 | Poly | 500 | 0.931 | 0.693 | 0.001 | SVC | 16261156 | Check mo… |
| 2021-07-12 14:48:50 | 1626115602 | Mihir | ipykerne | - | sklearn | 1.0 | Poly | 500 | 0.933 | 0.694 | 0.002 | SVC | 16261156 | Check mo… |
| 2021-07-12 14:48:01 | 1626115552 | Mihir | ipykerne | - | sklearn | 1.0 | Poly | 500 | 0.876 | 0.649 | 0.002 | SVC | 16261155 | Check mo… |
| 2021-07-09 16:27:58 | 1625002471 | Mihir | ipykerne | - | sklearn | 1.0 | Poly | 500 | 0.735 | 0.597 | 0.021 | SVC | 16258624 | Check mo… |
After clicking the "Compare" button, a table-like comparison between different runs will be generated, (as shown in Figure 5) allowing you to easily compare logged data across different runs. The parameters that differ across the runs are highlighted in yellow. This gives the user an idea of how model performance has changed over time based on the change in parameters.
Figure 5: Comparing logged runs in MLflow UI in detail
Figure 5 compares different logged runs in MLflow in detail. The tags, parameters and metrics are in different rows and the runs are in different columns. This allows a user to compare details of interest for different runs in a single window. The parameters which are different in different runs are highlighted in yellow. For example, in the experiments the parameters max_features and ngram_range were changed for different runs and hence they are highlighted in yellow in the image above.
Text in image:
Nlp_with_disaster_tweets > Comparing 5 Runs
| Run ID: | 7a1448a5f88147c093 c357d787dbe3 |
264533b107b04be3 bd4981560bad0397 |
7670578718b3477abb 798d7e404fed6c |
D2372d5873f2435c 94dc7e633a611889 |
Fdc8362b2f37432f9 a4128fa522d80cb |
|---|---|---|---|---|---|
| Run Name | 1626115725 | 1626115688 | 1626115602 | 1626115552 | 16265862471 |
| Start Time | 2021-07-12 14:48:54 | 2021-07-12 14:48:16 | 2021-07-12 14:48:50 | 2021-07-12 14:46:01 | 2021-07-09 16:27:58 |
| Parameters | |||||
| C | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
| Kernel | Poly | Poly | Poly | Poly | Poly |
| Max_features | 500 | 500 | 500 | 500 | 500 |
| Max_iter | -1 | -1 | -1 | -1 | -1 |
| Ngram_range | (1.3) | (1.2) | (1.1) | (1.1) | (1.1) |
| Norm | 12 | 12 | 12 | 12 | 12 |
| random_state | 42 | 42 | 42 | 42 | 42 |
| Metrics | |||||
| Mean f1-score-train | 0.93 | 0.931 | 0.933 | 0.876 | 0.735 |
| Mean f1-score-val | 0.694 | 0.693 | 0.694 | 0.649 | 0.597 |
| std f1-score-train | 0.001 | 0.001 | 0.002 | 0.002 | 0.021 |
| std f1-score-val | 0.008 | 0.009 | 0.01 | 0.013 | 0.042 |
Changes in the parameters and metrics across different runs can also be laid-out in a Scatter Plot. The values of the x-axis and y-axis can be set to any parameter or metric allowing the user to analyze the changes. In Figure 6, the reader can analyze the change in validation, in this case the mean F1-score, over different values for the parameter 'max_features'. If you hover over a data point, you can see details about that run.
Figure 6: Configuring the scatter plot to visualize the effects of different parameter configurations in the logged runs
A demonstration of MLflow's capability of plotting a graph using details from different runs. You can select a particular parameter on X-axis and a metric you want to monitor on the Y-axis; this will create a scatter plot with the details on the corresponding axis on the go and you will be able to visualize the effects of the parameter on the metric to get an idea about how the parameter is affecting the metric.
Text in image:
Scatter Plot
X-Axis: max_features
Y-Axis: mean F1-score-val
| Run Name | 1626115552 |
|---|---|
| Start Time | 2021-07-12 14:46:01 |
| C | 1.0 |
| Kernel | Poly |
| Max_features | 500 |
| Max_iter | -1 |
| Ngram_range | (1.1) |
| Norm | 12 |
| random_state | 42 |
| Mean f1-score-train | 0.876 |
| Mean f1-score-val | 0.649 |
| std f1-score-train | 0.002 |
| std f1-score-val | 0.013 |
The Parallel Coordinates Plot is also useful, as it shows the viewer the effect of the selected parameters on the desired metrics at a glance.
Figure 7: Configuring the parallel coordinates plot to visualize the effects of different parameters on the metrics of interest
In this image a parallel coordinates plot is configured. You can select different parameters and metrics using the provided input windows, based on which the parallel coordinates plot is updated. This plot can provide an idea about the results you get using different configurations in the experiments. It can help in comparing different configurations and selecting the parameters that perform better.
Text in image:
Scatter Plot – Contour Plot – Parallel Coordinates Plot
Paramters: random_state, norm, max_iter, max_features, C, kernel, ngram_range
Metrics: mean F1-score-val
| random_state | norm | Max_iter | Max_features | C | Kernel | ngram_range | Mean F1-score-val | |
|---|---|---|---|---|---|---|---|---|
| 46.20000 | -1.10000 | 500.00000 | 1.10000 | 0.69400 | ||||
| 46.0000 | -1.10000 | 500.00000 | 1.10000 | (1.3) | 0.68000 | 0.68 | ||
| 45.0000 | 450.00000 | |||||||
| 44.0000 | -1.05000 | 400.00000 | 1.05000 | 0.66000 | 0.66 | |||
| 43.0000 | 350.00000 | |||||||
| 42.0000 | -1.0000 | 300.00000 | 1.00000 | poly | (1.2) | 0.64000 | 0.64 | |
| 41.0000 | 250.00000 | |||||||
| 40.0000 | -0.95000 | 200.00000 | 0.95000 | 0.62000 | 0.62 | |||
| 39.0000 | 150.00000 | |||||||
| 38.0000 | -0.9000 | 100.00000 | 0.90000 | (1.1) | 0.60000 | 0.6 | ||
| 37.80000 | -0.90000 | 100.0000 | 0.9000 | 0.59700 |
There are other important points to note with MLflow tracking:
metrics.rmse < 1 and params.model='tree'. This is very helpful when you need to dig up a run with specific parameters executed in the past.Feel free to contact us at statcan.dsnfps-rsdfpf.statcan@statcan.gc.ca and let me know about other interesting features or use case you like to use that you feel could have been mentioned. We will also have an opportunity for you to Meet the Data Scientist to discuss MLFlow in greater detail. See below for more details.
If you have any questions about this article or would like to discuss this further, we invite you to our new Meet the Data Scientist presentation series where the author will be presenting this topic to DSN readers and members.
Tuesday, October 18
2:00 to 3:00 p.m. EDT
MS Teams – link will be provided to the registrants by email
Register for the Data Science Network's Meet the Data Scientist Presentation. We hope to see you there!