Extract data from flat advertisement web site and analyse it. This project has 4 parts:
- extract data using
requests, beatifulsoup - encoded travel time with
route API - present relations between
location, travel time, apartment's sizeandprice - develop
scikit-learn modeltopredict price
Both subfolders rental and sale contain strictly similar code with slightly different data source. The differences are only between files analyse.ipynb in sale and rental subfolders and are marked with red color or crossed text.
In this subfolder analysis is performed on rental flat offers (average prices in range 1500-5000 zł per flat).
In this subfolder analysis is performed on sale flat offers (average prices in range 250.000-1.000.000 zł per flat).
This subfolder uses pickled mashine learning algorithms, which can be learned by first running files analyse.ipynb from rental and sale subfolders. The goal of the algorithm is to predict rental or sale prices of apartaments. Interactive data typing is already supported.