Traffic Flow Forecasting

Armin Catovic^ex-Ericsson, armin.catovic@schibsted.com, armin.catovic@hotmail.se

This is the final project in KTH Microdegree, EP232U - Deep Neural Networks. Here we evaluate a number of different univariate forecasting models in the context of traffic flow prediction. For full explanations and code, please refer to the Jupyter notebook TrafficForecasting.ipynb.

We adapted the code (with some modifications/optimizations) from time-series-forecasting-with-python repo.

Pre-requisites

• Python >= 3.7
• Python dependencies as specificed in requirements.txt, i.e. pip install -f requirements.txt
• Internet connection since we load the dataset from another GitHub repo