Webscraping associated with a little spice of machine learning to find patterns on second-hand car prices. (on this particular case it was Fiat-Siena)
There is a snippet of code used for webscraping purposes collecting announced cars properties as price, mileage, year, etc.
As soon as you have enough data collected from any given car model, you can play with some data analysis, I tried to figure out any kind of pattern on the file carros.csv, and got some beautiful polynomial regression.
Using sklearn it was possible to find the "obvious", using extratrees classifier it was found that the main feature that affects a second-hand car price was the mileage.
All code found here was made for learning purposes, feel free to make use of anything here.