Embedded Files
Google Trends Malaria
Google Trends Malaria
Problem: Can we innovate new "proxy" measure for public health warnings?
Our idea: Use Google Trends data to draw pattern insights.
Learning (till now): Seasonal patterns can be seen for lot of infectious diseases.
GitHub - Python code
Predicting Diabetes
Predicting Diabetes
Creating models to predict diabetes using machine learning techniques and 4 waves of public DHS database for India with around 1 million individuals.
Insurance Demand
Insurance Demand
Problem: Consumers often buy sub-optimal insurance products.
Our idea: Created videos testing two behaviour change ideas - smart questions & smart advice.
Learning (till now): consumers change choice when equipped with Qs to ask.
Links: Funding of $40,000 from IIM-A
GitHub - Some Machine Leanring code
Health supply & demand
Health supply & demand
Problem: Who do mothers not consume all free tablets provided to them?
Our solution: Smart data insights using Indian Health Database (more than 1 million observations) and data from own survey.
Learning: Contrary to the existing belief that this is a demand side problem. I find that the majority of the gap that exists is explained by the imperfect supply of tablets.
Preferences & Well-being
Preferences & Well-being
Problem: Well-being measurment does not take consumer preferences into account.
Our solution: A new theoretically founded methedology.
Learning: Accounting for preferences changes poverty conclusions.
Links: Read full paper here
Budget Dashboard
Budget Dashboard
Collaborated with Centre for Budget Governance Accountability, India to create the first Indian State Budget dashboard.
Mathematical model
Mathematical model
Problem: How do we learn more about abortion behaviour?
Our idea: Created new structural statistical model that exploits data insight of differential probabilities of birth for boys and girls.
Learning: Women abort more than once in the same pregnancy.
Decision-making
Decision-making
Problem: What do we know about decision-making in multiple generation households?
Our idea: Quasi-experimental design with panel data to isolate impact.
Learning: Strong generation and gender divisions of power
Links: Read article here.
Behavioural change
Behavioural change
Problem: Why do people litter?
Our solution: Implemented a garbage disposal solution to change behaviour. Based on behavioural science and design methods.
Links: Read more on Bloomberg
Public goods
Public goods
Problem: Do people living in slums pay a poverty premium?
Learning: Residents often pay more than authorised colony residents for basic services. The highest premium is in the case of water with costs that are 4 to 7 times.
Method: Data sights of own survey.
Links: More information can be found in the paper and article.
School Ratings
School Ratings
Problem: How to drive quality in the affordable private schools market?
Our solution: Created a school assessment measure. Based on quantitative modeling.
Google Sites
Report abuse