C. Organization of Data for Analysis · Organization of Data for Analysis

Finding and Using Secondary Data: Statistics Canada and Other Databases

C2.5 C1.3About 70 minutes

Learning goals

I can find reliable secondary data from Statistics Canada and other well-organized databases.
I can evaluate a secondary source using its metadata: who collected the data, why, how, when and with what definitions.
I can read data-quality symbols and footnotes in published tables.
I can download, reorganize and clean secondary data so it is ready to analyse.
I can cite a data source correctly.

Words to know

Statistics Canada
Canada's national statistical office. It runs the Census of Population and many surveys, and publishes the results, mostly free online.

data table
A published table of aggregate data that can be filtered by geography, time period and characteristics, and downloaded.

reference period
The date or time span the data describe (for example, the 2021 census date, or July 2025). It is not the same as the release date.

open data
Data that governments or organizations publish free for anyone to use, usually under a licence that asks you to credit the source.

CSV file
Comma-separated values: a plain-text file in which each line is a record and commas separate the fields. Any spreadsheet can open it.

wide and long format
Wide format has one column per time period or group; long (tidy) format has one row per observation, with a column naming the period or group.

citation
A reference that tells readers exactly where data came from: the organization, the title or table, the date and how to find it.

You don't have to collect everything yourself

How many people in your town walk to work? How has the price of groceries changed since you were born? How many snowy days did your city have in 1995? You could never collect these data yourself, but someone already has.

The curriculum asks you to collect data from secondary sources, for example by using the Internet to access reliable data from a well-organized database, or by using print sources such as newspapers and magazines, and then to organize the data to answer a question. In this lesson you will learn where to look, how to judge what you find, and how to get it ready for analysis. Your Culminating Data Management Investigation will almost certainly use these skills.

Statistics Canada

Statistics Canada is Canada's national statistical office. Under the Statistics Act, it collects data about Canada's people, economy and environment, and it must keep individual responses confidential.

The Census of Population is held every five years and describes the whole country, down to small areas such as your neighbourhood.
Ongoing surveys produce regular statistics, such as the monthly Labour Force Survey (the source of Canada's unemployment rate) and the Consumer Price Index (the measure of inflation).
Results are published free in searchable data tables that you can filter by geography, time period and characteristics, then download as CSV files. Each table has a table number, and notes and definitions describe how the data were collected.
Some anonymized microdata files (Public Use Microdata Files) are available for detailed analysis.

When the curriculum was written (2007), Statistics Canada offered E-STAT, an interactive database for schools, and the curriculum names it as an example. Today, students use the same free data tables and census products on the Statistics Canada website.

Other reliable databases

Environment and Climate Change Canada: historical climate data (daily temperatures, precipitation, snow on the ground) for weather stations across Canada.
Open Government Portal (open.canada.ca) and the Ontario Data Catalogue (data.ontario.ca): thousands of federal and provincial data sets.
Municipal open data portals: many cities publish data on transit, trees, building permits and more.
Bank of Canada: exchange rates and interest rates.
Elections Canada and Elections Ontario: election results by riding.
Census at School: an international project (named in the curriculum) where students answer a class survey and the anonymous answers are shared as a database for classroom use.
First Nations Information Governance Centre (FNIGC): First Nations-led surveys such as the First Nations Regional Health Survey, collected and governed under the OCAP® principles you met in Designing Experiments and Ethical Data Collection.
Sports leagues and research articles for specialized topics.

Print and online news sources (newspapers and magazines) often report statistics. Treat them as a pointer: find the original source the article is based on and check it.

Working with secondary data: from a question to a cited, organized data set ready for analysis.

Is this source reliable? Read the metadata

Ask these questions about any secondary data set:

Who collected it? A national statistical agency, a university, a company selling a product, an advocacy group?
Why was it collected? Data collected to promote something may be selected or presented to support that goal.
How was it collected? Census or sample? How large? Random? What was the response rate? What questions were asked?
When? The reference period (when the data describe) and how current it is.
What exactly does each variable mean? "Household", "family", "full-time" and "urban" all have precise definitions that may differ from everyday use. Check units and geography (city or metropolitan area?).
Are there footnotes or quality symbols?

Statistics Canada tables use symbols such as these. A table can also show other symbols; the legend under each table explains them.

Quality symbols in a published table. These numbers are made up to show how the symbols work; always read the legend under a real table.

Worked example

Example 1: Choosing a source

Name a suitable secondary source for each question. (a) Has the mean January temperature in Thunder Bay changed since 1950? (b) What percentage of people in your census division speak both English and French? (c) How many trees does your city own on its streets? (d) What is the current exchange rate between the Canadian and US dollar?

(a) Long-term weather records: Environment and Climate Change Canada historical climate data (check whether the station moved over the years).
(b) Knowledge of official languages by area: the Census of Population (Statistics Canada).
(c) Many cities publish a street-tree inventory on their municipal open data portal.
(d) The Bank of Canada publishes daily exchange rates.

(a) ECCC climate data, (b) Statistics Canada census, (c) municipal open data, (d) Bank of Canada.

Worked example

Example 2: Reading a table with quality symbols

Using the illustration above: (a) Which towns' values can be used normally? (b) What should you write if you use Town B's value? (c) Why is Town D's value missing, and can you estimate it?

(a) Town A has no symbol, so its value can be used normally.
(b) Town B is marked E (use with caution): the estimate is less precise, often because few survey respondents were in that group. You can use it, but you should say so: "about 3150 (use with caution)".
(c) Town D is marked x: suppressed to protect confidentiality, usually because the count is so small that individuals might be identified. You should not try to work it out from other numbers in the table; treat it as unavailable.
Towns C (F) and E (..) have no usable value at all.

Only Town A is fully reliable; Town B can be used with a caution; Towns C, D and E have no usable value.

Worked example

Example 3: Checking a statistic in the news

A magazine article says, "A new survey shows Canadian teens spend 9 hours a day on screens!" Before using this number in a project, what should you do?

Find the original source. Who ran the survey? An app company, a health agency, a university?
Check the method: how many teens, how were they chosen (random or an online opt-in poll?), and what was the exact question? Does "screens" include school work and TV in the background?
Check the definitions and date: which ages count as "teens"? When was the survey done?
Compare with another reliable source, such as a national health survey.
If you cannot find or check the original source, do not use the number, or use it only with a clear warning.

Trace the statistic to its original source, check who, how, what and when, and compare it with another reliable source before trusting it.

Worked example

Example 4: Reorganizing a downloaded table

A downloaded table (made-up numbers) is in wide format: columns Province, 2023, 2024, 2025, with rows Ontario 41.2, 41.8, 42.5 and Manitoba 38.0, 38.6, 39.1 (an index value). Reorganize it into long (tidy) format.

Tidy data have one row per observation. Here an observation is one province in one year, with one value.
Create three columns: Province, Year, Value.
Each wide row becomes three long rows: (Ontario, 2023, 41.2), (Ontario, 2024, 41.8), (Ontario, 2025, 42.5), and the same for Manitoba.
Now you can sort, filter and graph by year or province, and add more years by adding rows, not columns.

Long format: 6 rows × 3 columns (Province, Year, Value). This is easier to filter, sort and graph in a spreadsheet.

How to cite a data source

Include: the organization (author), the title of the table or data set and its number or identifier, the release or publication date, the web address (or DOI), and the date you accessed it. Statistics Canada shows a ready-made citation ("How to cite") on each table page. Credit open data according to its licence.

Just for fun

Statistics Canada counts everyone in the country, yet nobody ever asks it, "How are you doing?" Luckily it already has a survey for that.

Hands-on activity

Download, organize and cite a real data set

You need: a computer with Internet access and a free spreadsheet program.

Choose one question, for example: "How did my province's population change over the last 10 years?" or "How cold was each day last January in my town?"
Find the data: for population, search the Statistics Canada website for population estimates for your province; for weather, use Environment and Climate Change Canada's historical climate data and choose the station nearest you.
Before downloading, read the notes or metadata: what is the reference period, what are the units, and are there any quality symbols?
Download the CSV file and open it in your spreadsheet. Keep the original file unchanged; work on a copy.
Organize the copy: delete rows and columns you do not need, give clear field names with units, put it in long (tidy) format if needed, and check for missing values.
Calculate one useful summary (for example the total change, or the mean daily temperature) and make one graph.
Write a full citation for your source and two sentences about its reliability.

What to look for: a clean, clearly labelled table; a correct summary; a citation with organization, title or table number, date and web address; and a reliability comment such as "collected by a national agency using a consistent method; values for the most recent year are preliminary and may be revised."

Your turn

Name Canada's national statistical agency and one of its regular surveys.

What is the difference between the reference period and the release date of a data set?

A table value is marked F. What does it mean and what should you do?

Why should you find the original source of a statistic you read in a newspaper?

Give two questions to ask about who collected a data set and why.

Convert this wide table to long format: Town, 2024, 2025; Elmvale 3, 5; Lakeview 4, 2.

More practice: warm-up

What does CSV stand for?

Which federal department provides historical weather data?

In a Statistics Canada table, what does the symbol x mean?

Should you edit the original downloaded file?

Name two things a data citation should include.

More practice: on your own

A company that sells sleep supplements publishes a survey showing that 80% of Canadians sleep badly. List three concerns about this source.

You compare city populations from two sources. One says 1 200 000 and the other 2 400 000. Suggest a reason based on definitions.

Explain why Statistics Canada suppresses some small counts.

A downloaded climate CSV has a column of daily maximum temperatures with some cells marked M. What should you do before calculating the monthly mean?

A newspaper reports that "crime rose 50%" in a small town. What secondary data would you look for to judge this claim?

Challenge

Plan a short investigation using only secondary data to answer: "Do Ontario regions with older populations have a higher share of people who walk to work?" Name the data needed, the source, the organization steps and one limitation.

Two tables use different geographic codes for the same areas. Explain why joining them by area name instead of by code can go wrong.

A data set's values for the most recent year are labelled "preliminary". Explain what this means for your analysis and how you would report it.

Check your understanding

Quiz

Which organization runs Canada's Census of Population?

In a Statistics Canada table, the symbol E means:

A magazine quotes a surprising statistic. What should you do first before using it?

Which question about a data set is about its reference period?

Why is "long" (tidy) format often better than "wide" format in a spreadsheet?

A data set about First Nations communities is governed by the OCAP® principles. This means that:

For parents and teachers

Big idea (C2.5, C1.3): reliable secondary data are available free from Statistics Canada and other agencies. Students must judge a source's reliability (who, why, how, when, definitions), read quality notes, organize the downloaded data, and cite the source.

How to help: Explore the Statistics Canada website or your city's open data portal together and find a number about your own community. When you see a statistic in the news, ask where it came from and look up the original.

Common difficulties:

Trusting a number because it appears online or in print, without checking the original source.
Ignoring footnotes, quality symbols and definitions.
Confusing the reference period with the release date.
Editing the original download instead of a copy, or forgetting to record where the data came from.
Joining tables by names rather than by unique codes.