NJUPT Big Team Project for Software Engineering

Jupyter-powered Customer Classification Software.

Team list:

F18030115 Alladin Al-Gannam

F19030132 Anna Akulova

Features

Simple and easy to use professional software for business
Beautiful visualizations based on your csv data
Import, save and even export results as PDF
Web-based application
Using bleeding-edge technologies in AI

Tech

Our project uses a number of open source projects to work properly:

Python - easy and fast programing language
Jupyter-notebook - awesome web-based environment
Tensorflow - end-to-end open source platform for machine learning

Installation

The project requires:

numpy==1.20.1
pandas
tensorflow
keras
sklearn
pycaret
seaborn
matploitlib
scikit-optimize

Clone this repository:

git clone https://github.com/hetake2/njupt-teamproject-2021
cd njupt-teamproject-2021/

Run pip to install libs:

pip install -r requirements.txt

Start the app by script:

sh start.sh

How-to-use

In order to use this app, simply edit Data_Insurance_TGI.csv file and run the program to see results. For users we recommend to run all sections before "4. Transform & Compare Models in PyCaret".

For Devs

Thanks to the Jupyter framework, we are able to store and edit code directly in our program, so, you can feel free to edit it and instantly get the results.

Starting from "4. Transform & Compare Models in PyCaret" section, you can train your own model for this project, but you have to use an Nvidia GPU and Cuda drivers.

License

GNU GPL 3

Free Software, Hell Yeah!