https://7daysofcode.io/matricula/data-science
- Data clean up
The first challenge involves data clean up. Which is, getting some, or as I did, all, csv's and cleaning the data on them.
For that I have created the file 'data_cleanup.ipynb' where I gather all csv's, clean the data, and output a new csv file containing all cleaned up data from all files.
If you want to use the output file alone, it is called 'despesa_ceaps_2008_2022_formatted.csv'.
- Story Telling
The second challenge involves story telling, where we try to tell a story with the data in the most engaging way possible.
Here I used the cleaned up data from CEAPS to tell a story about the data. I chose to analyse a certain characteristic on the dataset and dived in it, investigating patterns and possible explanations to such behaviors.
- Forecasting
The third challenge involved forecasting over the CEAPS data. Here I tried to predict future senators' expenses based on their history using Facebook Prophet algorithm.
- Recommendation Models
The fourth challenge is about creating recommendation models. Moving away from CEAPS data, this challenge used the MovieLens 100K dataset, where we have available 100k movie ratings from 1k users on 1.7k movies.
The idea is to try some different approaches for recommending movies to people, like, randonly, most-popular, genre preferences and possibly getting to a ML model to consider multiple factors.