Summary of results may be found here: https://docs.google.com/document/d/17UUAweMWBC5K7yGNm9FLFASn2LdhuHec4Ew-loJSExs/edit?usp=sharing

To install this package you will need the 'conda', or another virtual environment. Conda is the one that I used which now has the name miniconda and is downloadable here https://conda.io/miniconda.html conda create --name=hprice python=2.7.13 conda activate hprice

• From this directory install all relevant packages: pip install -r requirements.txt conda instal jupyter
• Complile code: python setup.py develop

Non-standard packages installed are: https://github.com/scikit-learn-contrib/sklearn-pandas

Given a csv file, with housing prices that we wish to predict, we use the full command: predict_pricesi -i data/single_family_home_prices_to_predict.csv or -i < your file> that is in the same format as the

Please feel free to look at the improved_pipeline.ipynb, though realize that it is an exploration-style notebook and not intended for presentation purposes.