Data Science for Social Impact
Using data to work on problems (poverty, health, justice) and the traps that make 'data for good' go wrong.
AI tutor is turned off for this class
Use the DATASCI154 lectures, notes, and assignments below to keep learning.
Lesson overview
Using data to work on problems (poverty, health, justice) and the traps that make 'data for good' go wrong.
Teacher script · 45 min
- 0–5
Hook
Data science for social impact starts with a humane question (who lacks clean water, which neighborhoods get over-policed) and uses numbers to see it clearly. Done right, data makes invisible suffering visible and undeniable. Done carelessly, it hides people inside averages.
- 5–15
Explore
Students do the activity in pairs: Find a statistic in the news. Ask: how was it measured, and who might be uncounted? Often the gap changes the whole story.
- 15–30
Explain
Data can show two things move together (ice cream sales and drownings) without one causing the other (summer causes both). In social work this matters enormously: act on a false cause and you hurt real people. Worse, a deployed model can change behavior and entrench the very inequity it 'predicted.'
- 30–40
Connect to the summit
Read aloud the "From the summit" section below. Ask students which ideas from today they recognize in it.
- 40–45
Check
Run the formative check below. Anyone who can explain a key term in their own words has it.
Student activity
Find a statistic in the news. Ask: how was it measured, and who might be uncounted? Often the gap changes the whole story.
Slides
Formative check
- 1.In your own words, what is "Selection bias"? (Looking for: When the data you collected isn't representative of who you care about.)
- 2.In your own words, what is "Correlation vs causation"? (Looking for: Things moving together doesn't mean one causes the other.)
- 3.In your own words, what is "Proxy variable"? (Looking for: A measurable stand-in for something you can't measure directly, and a common source of bias.)
Carry-away concepts
- Selection bias
- When the data you collected isn't representative of who you care about.
- Correlation vs causation
- Things moving together doesn't mean one causes the other.
- Proxy variable
- A measurable stand-in for something you can't measure directly, and a common source of bias.
- Feedback loop
- When acting on a prediction changes the world to confirm it.
From the summit · the Stanford source
You apply data science to social problems responsibly, learning the statistics and the ethics of measuring human lives.
This module is based on a Stanford course.