Sixteen neural networks - eight for each of the advanced and basic calculator. Currently trained on data from 778 000 homes across England and Wales, and tested on data from 67 000 different homes. Uses self-made neural network class, which has four functions: query, train, save and load.

Train advanced calculator and save weights using NN_AdvTrain.py. Change path of training data on line 30. Rename path/names of weights on lines 96 - 103, as desired. Train basic calculator and save weights using NN_BasicTrain.py. Change path of training data on line 30. Rename path/names of weights on lines 90 - 97, as desired.

Test calculators with NN_AdvTest.py and NN_BasicTest.py. Load weights of neural networks with previously set path and filename. Edit filepath of testing data on line 38 of both calculators.

Webapp runs through main.py. Running main.py will start a development server on local host, which can be accessed with http://127.0.0.1:5000/ . Deployed web app can be found here: yourgreenhome.nw.r.appspot.com OR yourgreenhome.appspot.com

Enter pip install -r requirements.txt in console.