PawsHelper

Authors (all members contribute equally)

Chen Wang, Chujie Chen, Tongshu Yang, Xinyu Zhang, Xueting Feng, Zhe Zhang

DESCRIPTION

This repo contains four main parts: data manipulation, model training, React.js front end, and Flask back end.

Main Dependencies

• Data
  • Python 3.7
  • Jupyter-notebook
  • Numpy/Scipy
  • Pandas
  • Seaborn
• Model
  • Scikit-learn
  • Matplotlib
• Front end
  • React
  • Nivo
• Back end
  • Node.js
  • Flask
  • Gunicorn

File Structures

/data

It contains all raw datasets and cleaned datasets.

/examples

It contains all jupyter notebooks (data join, data cleaning, model training and evaluation).

/src

It contains scripts for

• /src/.

  • Data join
  • Data cleaning
  • Model training

• /src/ui

  • main interactive data visulazation front end with React.js

• /src/flask

  • predictor with Flask backend and simple front end /tests It contains testing scripts

INSTALLATION

1. git clone this repo
2. cd to /src/ui and npm i to install all required dependencies
3. create a pythonic virtual enviorment and install all required python libaries

EXECUTION

https://k31ex.csb.app/

We highly recommend you to use the url above to have an eye candy. But if you are interested, you can still test everything locally.

For scripts under /src/., it's likely you don't have enough computational power / space to finish running those .py code. We suggest you to upload notebooks under /examples to google colab and test on the clound.

For the React-Nivo UI, if you are familiar with NodeJS and React.js, you follow the /src/ui/README.md to test the UI locally.

For the Flask, you can simply run python app.py to launch the Flask predictor app.