GetThereGreen is an interactive simulation where users can see how their commute to work affects the quality of the air around them.
The API is hosted on a Raspberry Pi. This runs NGINX, Django, and TensorFlow in order to return predictions back to the frontend. The frontend is made in React/Typescript and hosted on Firebase
We used Tensorflow/Keras in order to create 4 highly accurate model (Multi-Layer Perceptron) that predicts the concentration of Sulfur Dioxide, Nitrogen Dioxide, Carbon Monoxide, and Surface Level Ozone. The input data that wes used was acquired from the US Census data database under the specific header of B08301. The air quality data was acquired from the EPA (United States Environmental Protection Agency) and is used as the labels for our training data.
The React code is stored in the app/ directory. You will find the components and such in app/src, while metadata and other information is stored in public/src
The API folder is an API (Application Programming Interface) was created using the DJANGO Web Framework. Machine Learning models are stored in the api/models folder where all the models have been zipped in order to save data for the repository. Most of the backend code that is used to process the incoming POST request is written in the api/prediction/views.py file. In order to run the following api, do the following:
1. Install Dependencies:
pip3 install -r requirements.txt
2. Make Migrations:
python3 manage.py makemigrations
python3 manage.py migrate
In the Network folder, the models were trained and the data was cleaned. Firstly, the census.py file located in network is used in order to clean the data that is used to be fed into the Machine Learning algorithm. The main.py file is the testing file in whcih the models were created and functions that are used to format the predictions of the neural network were made.