This is a small project which collects 10 day weather forecast information from the weather underground API and plots these predictions against the actual temperature. This data is plotted via a flask app. Data and forecasts are all for the Minneapolis/St. Paul airport. Only the temperature forecast is recorded and compared.

Python 2, matplotlib, mpld3, flask and numpy are required.

fd.py and wd.py are Python programs which periodically query the weather underground API. The temperature forecasts are stored in the data/forecast.*.txt files and the actual temperatures are stored in data/data.txt. The format for these files is lines of time-stamps and temperature in Fahrenheit.

To use fd.py and wd.py a weather underground API key is needed. This should be stored as a string in the environmental variable KEY.

flaskapp.py is a flask app which plots the forecasts against the actual data.

giving the command: $ python flaskapp.py should start a local web-server and you can see the results on the following pages:

127.0.0.1:5000/png serves a png file which shows the actual temperature in black and the forecast temperatures colored according to how old the forecasts are. Blue forecasts were made recently and red forecasts were made 10 days ago.

An example of the data:

A zoomed in example:

127.0.0.1:5000/d3/ is a rendering of the data using d3. It allows interactive zooming and panning but is slow for large data sets.

show_data.py plots the data using matplotlib.