Unlocked Phone Reviews amazon

clone this repo

git clone https://github.com/abdoufermat5/text-mining-reviews

Simple way to run the app

If you have docker and docker compose installed, you can run the app with the following command:

docker compose up --build -d

Then go to http://localhost:8501

if not started rerun the command again

If you don't have docker, you can run the app with the following steps:

install dependencies

bash pip install -r requirements.txt

download the dataset from kaggle

PS: You don't really need to do this step, because the dataset is already stored in a mongoDB cluster

https://www.kaggle.com/datasets/PromptCloudHQ/amazon-reviews-unlocked-mobile-phones

Then move the csv file to the data/ folder

notebooks

• notebooks/stats_on_other_features.ipynb

here we explore the other features of the dataset

• notebooks/preprocessing.ipynb

here we preprocess the data

• notebooks/model.ipynb

here we build different models and compare them

streamlit app

We've built a streamlit app to make it easier to understand the data and the models

To run the streamlit app

go to the deploy folder cd deploy

install local package fermat-helpers pip install .

run the app bash streamlit run Accueil.py

go to the url displayed in the terminal http://localhost:8501

Some screenshots of the app

• Visualisation of the data and Stats

• Preprocessing page

• Exploring the models

• Predictions