While looking for jobs, I came across a data analysis practice problem here. Given household power usages, which households which own electric vehicles, and when those vehicles are charging, the goal is to:
- Predict which households own electric vehicles
- Predict when those households are charging
- Find anything else interesting in the dataset
This directory contains my solution. Overall, I spent about 2 weeks on this problem, with about 30-40 hours each week. The first week I played around with the data, made plots, and gained intuition for the problem as a whole. I also taught myself the pandas library and parts of the scikit-learn library, which were significantly helpful in terms of productivity. The second week I coded the solutions and wrote the ipython notebooks.
Several pdfs can be found in the "docs/" directory. A summary of the results and methods can be found in the file named "writeup.pdf". The detailed reasoning and code can be found in each of the ipython notebooks (and their pdf versions in the "docs/" directory), with some helper code in "util.py" and "markov.py".
If there are any questions or comments about the analysis, I can be reached at scottmsul at gmail.com. Thanks for reading!