This contains my solution to a task called Rebel Rescue.
- The assignment is to predict position of agents in space-time given extremely scattered information about their whereabouts.
- The most pertinent exercise hence was to wrangle the data in such as way as to enable carrying out estimations or learning models.
- My solution involved multiple imputation (Amelia), network algorithms (networkx), and Pandas operations.
- Because multiple imputation is involved, I became aware of the need to help python users to implement Amelia (a package for R, not python): the only software for handling missing data in cross-sectional, time-series data structures.