CarsDataset by Martin Maikov and Joosep Tiger Tilgar

At the moment the main conclusions regarding cars performance is based solely on technical specifications and car reviewers interpretation of those metrics, but there are other more interesting ways of approaching this matter. For example, out in the world there are up to hundreds of thousands of different generations of different car models from many different brands and a variety of makers. Therefore we have a lot of close to indistinguishable metrics to use to find interesting patterns and extract valuable data which can’t be gathered just by looking at the technical specifications themselves, but rather by analysing it with machine learning algorithms.

Code

Code is contained in .ipynb jupyter notebooks where one is for cleaning the dataset and other is for analysing the said dataset. The code for the other dataset analysis is on the other branch 'joosep-dev'. The code should work as is by just running it.