Project information

  • Category: Regression problem
  • Goal: Predict the value of a flat
  • Project date: April, 2020
  • Project URL: See the codes on Github

What is this project about ?

Thanks to a dataset collected on the French government website data.gouv.fr, I decided to try predicting the value of a flat in the North of France. In the dataset, we have values like the surface, the localisation, the number of rooms, and a lot of other information. I decided to work only on the flat because it was the most interesting for me because I am currently looking for an apartment. At the end of this project, I managed to have a RMSE (Rooted Mean Squared Error) of approximately 50k which is quite good. The model manage to predict very precisely most of the datas but make some huge errors on some of them, which increase the RMSE. To have an idea, I had 55% of the prediction correct at more or less 20k and 45% correct at more or less 10k. If you want to discover how I solved this problem, please don't hesite to click on the link and have a look at my codes on GitHub !