Data Science for Social Good Problem Templates

Presenter: Rayid Ghani

This workshop will describe 6 templates most data science problems in social good fall in to. These include:

• Early Warning: Can I detect which children are going to get lead poisoning early?
• Resource Prioritization: Can I prioritize homes to inspect with potential health and safety violations?
• Scheduling and Assignment: Can I determine which medics/ambulances/fire trucks to schedule/assign to respond to an incident?
• Routing: Can I determine which office/department to route an incoming service request/311 call to?
• Policy Recommendations: Can I recommend which policies to modify to improve maternal mortality?
• Impact Evaluation: Can I evaluate the impact of a new policy?
• Data Creation: Can I turn this text/images/video/audio I have collected in to “data” i can analyze?

We will work in small groups to identify problems in governments and non profits and interactively group them into these templates. For each template, we’ll go over how to define the problem, what data is needed, what analysis needs to be done, and how to validate the analysis.

Level:  All

Audience :

• People in governments, non-profits and corporations interested in understanding  social good problems where data can help.
• Data Scientists interested in helping governments and non-profits