This project focus on building a Neural Network to predict the number of bikeshare users on a given day. Imagine yourself owning a bikesharing company. You want to predict how many bycles you need because if they are too few, you will loose money from potential rides, if they are too many, you will waste money on buying cycles just settled around. You need to predict from historical datas how money bycles you will need to buy in the future.
Libraries like Tensorflow are built around a computational graph. This project shows how to build your own deep-learning library.
- Download anaconda
- Create a new conda environment:
conda create --name bikesharing python=3
- Activate the source
source activate bikesharing
- Ensure you have numpy, matplotlib, pandas, and jupyter notebook installed by doing the following:
conda install numpy matplotlib pandas jupyter notebook
- Run the following to open up the notebook:
jupyter notebook bikesharing-neural-network.ipynb
You can open the jupyter notebook directly in github bikesharing-neural-network.ipynb