D. Data · Data, Statistics and Mathematical Modelling

Big Data All Around Us: Collecting, Storing and Using Data

D1.1 A2About 65 minutes

Learning goals

I can identify a current context that involves a large amount of data.
I can describe how data is collected, stored, represented and used.
I can explain benefits, risks and consequences of large data sets, including privacy, security and bias.
I can explain what data sovereignty means and name the OCAP® principles.

Words to know

data set
A collection of values (numbers or categories) about a group of people, objects or events.

big data
Data sets so large that they must be stored and analysed with technology.

privacy
A person's right to control information about themselves.

data breach
When stored data is stolen, leaked or seen by people who should not have it.

consent
Permission given freely, knowing what you are agreeing to.

re-identification
Working out who a person is from data that had names removed.

implication
A possible effect that follows from a choice or situation.

data sovereignty
A community's right to govern data about its own people, lands and resources.

Before breakfast, you were already data

Think about a normal school morning. Your phone may have counted your steps while you slept. A weather app checked the temperature where you live. When you tap a transit card on the bus, a computer records the time and the stop. When you stream a song, the app records what you played and for how long.

None of this is a single number in a notebook. It is millions of records, collected automatically, from millions of people, every day. In this lesson you will learn to look at a context like this the way a mathematician does: what data is collected, how it is stored, how it is represented, how it is used, and what could go right or wrong.

Remember from Grade 8

In Grade 8 you sorted data into one-variable and two-variable data, collected continuous data, made scatter plots, spotted misleading graphs and built infographics. In Grade 9 you will work with much larger data sets, analyse them with new tools (quartiles, box plots, regression), and use data to build mathematical models of real situations.

What counts as a large amount of data?

A data set is a collection of values about a group of people, objects or events. A large data set has so many values that you could never look at them one at a time. You need technology (spreadsheets, databases and graphing tools) to organize and summarize it. People often call this big data.

Some Canadian examples:

The census. Statistics Canada runs a census of the whole country every five years. It counts people and households and asks about things such as age, language and housing.
Weather and climate. Weather stations across Canada record temperature, precipitation and wind many times a day, year after year.
Transit cards. Every tap of a fare card records a time, a place and a card number.
Health records. Hospitals and clinics store information about visits, tests and vaccinations.
Store loyalty programs. Every purchase with a points card records what was bought, where and when.
Apps and games. Apps record what you tap, how long you stay and often your location.

Each step up is 1 000 times bigger than the one before (the sizes are rounded, and computers sometimes use 1 024 instead of 1 000). A large organization can store petabytes: that is a million gigabytes.

The four stages: collect, store, represent, use

Every large data set goes through four stages. At each stage, ask your own questions.

The data life cycle. Data is collected, stored, represented (in tables, graphs, maps or dashboards) and used to make decisions. Those decisions often lead to collecting more data, so the cycle starts again.

Stage	Questions to ask	Possible concerns
Collect	What is collected? From whom? Did people agree (consent)? Is it collected automatically or by survey?	Collecting more than needed; people not knowing; some groups left out
Store	Where is it kept? Who can see it? Is it protected? How long is it kept?	Data breaches (hacking or leaks); data kept forever; data sold or shared
Represent	Is it shown as totals, averages, maps, graphs? Is anyone identifiable?	Misleading graphs; small groups becoming identifiable; important groups hidden in an average
Use	What decisions are made with it? Who benefits? Who might be harmed?	Unfair decisions if the data is biased; using data for a purpose people never agreed to

A checklist you can use for any large data set.

Implications and consequences

An implication is a possible effect that follows from a choice. A consequence is what actually happens as a result. Large data sets can have benefits and risks, often at the same time.

Possible benefits

Transit agencies can add buses at the busiest times.
Doctors and researchers can spot health trends sooner.
Weather data helps forecasters warn people about storms and extreme cold.
Census data helps governments decide where to build schools, hospitals and roads.

Possible risks

Privacy. Location data can reveal where someone lives, goes to school or worships.
Security. Stored data can be stolen in a breach.
Bias. If some people are missing from the data (for example, people without smartphones), decisions based on it may not serve them well.
Re-identification. Removing names is not always enough. Combining data sets (postal code + birthday + school, for example) can sometimes point to one person.
Representation. A graph or a single average can hide important differences between groups.

In Canada, laws set rules for how organizations may collect, use and share personal information. For example, the federal law called PIPEDA applies to many private-sector businesses.

Indigenous data sovereignty

Data sovereignty means that a community has the right to govern data about itself. Many First Nations follow the OCAP® principles, developed by the First Nations Information Governance Centre: Ownership, Control, Access and Possession. They say that First Nations communities own and control data about their people, land and resources, and decide how it is collected, stored, shared and used. Inuit and Métis organizations have also set out their own principles for research and data about their communities. When you evaluate a data set, ask: whose data is this, and who gets to decide how it is used?

Worked example

Example 1: A city transit fare card

A city's transit system uses tap-on fare cards. Describe the data, then describe possible implications and consequences of its collection, storage, representation and use.

What is collected? For each tap: the card number, the date and time, the stop or station, and the route. Some cards are registered to a person's name and email.
How big is it? If 400 000 riders tap twice a day, that is 400 000 × 2 = 800 000 records per day, or about 800 000 × 365 = 292 000 000 records per year. That is a large data set.
Storage: on the transit agency's servers. Questions: Who can see it? Is it encrypted? How long is it kept?
Representation: maps of the busiest stops, graphs of riders by hour, dashboards for planners. Totals by stop and hour do not show individuals.
Use and benefits: add buses at rush hour, plan new routes, find stops that need shelters.
Risks: a registered card shows one person's daily pattern (home, school, work). If the data is leaked or shared without consent, it could put someone at risk. Riders who pay cash are not in the data, so their needs could be missed.

The data helps plan better service, but it should be stored securely, kept only as long as needed, represented as totals rather than individual trips, and used only for the purpose riders agreed to. Planners should also remember who is missing from it (cash riders).

Worked example

Example 2: A fitness app

A free fitness app records steps, heart rate, sleep and GPS routes. The terms of service say the company may share "de-identified data" with partners. What are the implications?

Collection: automatic, every few seconds, from a phone or watch. The user agreed by accepting the terms, but most people do not read them.
Storage: on company servers, possibly in another country.
Representation: charts for the user; the company may publish maps of popular running routes.
Use: helps users set goals; helps the company improve the app or earn money from partners.
Risk: GPS routes that start and end at the same place every day can reveal a home address, even with the name removed. A public route map could show where people regularly exercise.

Benefits for the user, but real privacy risks. A careful user can turn off location sharing, make their profile private, and read what "de-identified" means in the terms.

Just for fun

A data scientist was asked to describe her weekend. She said, "Mostly sunny, 3 outliers, and a strong positive correlation between naps and happiness."

Hands-on activity

Data audit of your day

You need: paper and a pencil, and a grown-up to help with the phone or tablet settings.

Make a table with four columns: Device or service, Data it collects, Who stores it, What it is used for.
Think through one day. Add every app, card, game, website, smart speaker or school system you used. Aim for at least 6 rows.
For each row, guess the data collected (location? time? what you bought? what you watched?).
With a grown-up, open the privacy or location settings on one app. Is location always on, only while using, or off? Change anything you both agree should change.
Pick the row that worries you most and the row that helps you most. Write one sentence for each explaining why.

What to look for: at least 6 rows, specific data types (not just "stuff"), and a sentence that names a real benefit and a real risk.

Your turn

A grocery chain's loyalty card records every item you buy. Name two pieces of data it collects with each purchase.

Give one benefit to the store and one benefit to the customer of the loyalty data.

Give one risk of storing loyalty data for many years.

A school board's attendance system records each student's absences. Is this a large data set? Explain.

What does re-identification mean?

What do the letters in OCAP® stand for?

A news site shows a map of a city with one dot per home where a parcel was delivered last week. What is the problem with this representation?

More practice: warm-up

Name the four stages of the data life cycle in order.

Which Canadian agency runs the national census?

How often is the Canadian census held?

How many megabytes are in 1 gigabyte (using 1 000)?

True or false: removing names from a data set always makes it impossible to identify people.

More practice: on your own

A smart thermostat records the temperature inside a home every 5 minutes. How many readings does it record in one year (365 days)?

Using the thermostat in the last question, what could the data reveal about the people in the home? Why might that matter?

A survey about bus service is only posted online. Whose views might be missing, and what could be a consequence?

A hospital publishes the average wait time in its emergency department as 3 hours. Why might this single number be a weak representation?

A First Nation is asked by a university to share health survey data about its members. Using OCAP® principles, name two things the community would expect to decide.

Challenge

A streaming service has 2 million Canadian users. Each user streams 12 songs a day on average, and each play creates a record of about 500 bytes. About how much data is created per year, in gigabytes? (Use 1 GB = 1 000 000 000 bytes.)

A city wants to use data from a navigation app to decide which streets need new sidewalks. Describe one way the data could be biased and one way to reduce that bias.

Write a short paragraph (3 or 4 sentences) evaluating this statement: "If you have nothing to hide, it doesn't matter who collects your data."

Check your understanding

Quiz

Which is the best example of a large data set?

What are the four stages of the data life cycle used in this lesson?

A data set has names removed but still includes postal code, birth date and school. What is the main risk?

In OCAP®, what does the C stand for?

A survey about a new skate park is only shared on one social media app. What is the most likely problem?

A smart watch records heart rate once every minute. About how many readings is that per day?

For parents and teachers

Big idea: Expectation D1.1 asks students to identify a current context involving a large amount of data and describe potential implications and consequences of its collection, storage, representation and use. There is no single right answer: look for specific data types, realistic benefits and risks, and reasons.

How to help: Do the data audit together. Talk about one app your teen uses daily and look at its privacy settings with them. Ask "Who benefits from this data? Who could be harmed?" at the dinner table when a news story mentions data or a breach.

Common difficulties:

Answers that are too general ("they collect info"). Push for which info.
Thinking anonymous always means safe. Discuss re-identification.
Only seeing risks or only benefits. Most data contexts have both.

The arithmetic in this lesson (records per day or year) is a good chance to practise estimating with large numbers.