An example workflow using real estate data, creating forecasts and embedding them in an interactive map. This could be a starting point of a deeper analysis.
- The markers display forecasts of PHIV(Percentage of Homes Increasing in Value) for different zipcodes in the Chicago area. PHIV is one among many metrics available from Zillow Research Data on Quandl.
- The forecasts were generated using Prophet, it's intuitive interface make it a convenient choice to do time-series analysis.
- Altair makes it possible to generate vega-lite charts depicting forecasts.
- These vega-lite charts are then embedded into an interactive map using Folium.
Please have a look at the notebook for code used to generate the map.
Install Anaconda(https://docs.anaconda.com/anaconda/install/) and setup environment using:
conda create -n myenv python=3.4pip install -r requirements.txt
For accessing Quandl data an api key is needed: https://blog.quandl.com/getting-started-with-the-quandl-api
Prophet: https://facebookincubator.github.io/prophet/
Geopy: https://github.com/geopy/geopy
Altair and Vega-lite: https://altair-viz.github.io/index.html https://vega.github.io/vega-lite/
Folium: https://github.com/python-visualization/folium
https://www.dataquest.io/blog/data-science-portfolio-project/