This system is a Neural Network for pollution prediction of Mexico City

Tensorflow

pip installl tensorflow

or

pip installl tensorflow-gpu

pip install pandas
pip install -U scikit-learn
pip install seaborn

conda install -c jjhelmus tensorflow=0.12.0rc0
conda install -c anaconda tensorflow-gpu
conda install -c anaconda pandas=0.20.1
conda install -c anaconda scikit-learn=0.18.1
conda install seaborn

This code will generate CSV files from NetCDFs of a metheorological forecast form the IOA group at the UNAM. It should also work for other WRF model outputs. python NetCDF/makeCsv.py

File involved:

• NetCDF/NewBBOX.py
• NetCDF/Dominio.py

This code will generate a CSV file containing the Mexican holidays with these numbers: 0 = laboral, 1 = oficial, 2 = sep python Utilites/makeDayCsv.py

Input file:

• Data/Festivos.csv

Output file:

• Data/Festivos2019Merged.csv

python saveData.py

File involved:

• Utilites/FormatData.py
• Utilites/Utilites.py

4. Neural network training

auto.py

File involved:

• NNSystem/neuralNetworkGpu.py
• NNSystem/neuralNetwork.py
• NNSystem/neuralNetworkGpuMax.py

5. Neural network prediction

testPediction.py

File involved:

• prediction.py