- 1 – very happy
- 2 – rather happy
- 3 – not very happy
- 4 – not at all happy
Links:
- iPython Notebook code
- Web App to take survey and make happiness prediction
- ArcGIS Online Web Map
- Tableau Dashboard
Data was used to:
- Find features that correlate most with happiness.
- Create model to predict subjective, self-reported happiness levels for individuals and averaged for countries.
- Evaluate happiness levels in countries across the world to see spatial influence.
The questionaire includes hundreds of other questions, yet some topics (e.g. health) are more closely linked to happiness.The dozen questions with the highest correlation to happiness were used to make a classification model to predict a person's happiness category (which can then be grouped for average country happiness). Various algorithms were investigated including random forest, decision tree, and nearest neighbor.