Smart_Grievances_Redressal is a dashboard made by Team Scooptroop for monitoring people complaints against posted in various social media handles, like twitter as well as the grievances portal. .
- Dash - Main server and interactive components
- Plotly Python - Used to create the interactive plots
- Google Translate - Translate marathi/hindi or code mixed language into english
- Amazon transcribe - To convert speech to text, yet to be integrated with frontend
- We have also facilitated the use of amazon cloud 9 along with EC2 instance and lambda function generator to facilitate this feature as an independent API, which is able to handle large and live inputs and produce the output
- Classifies complaints without human intervention according to the department
- Text classification using deep learning approaches such as CNN and LSTM
- Modular use of Twitter API to get the most relevant data
- Code-mixed laguage could also be used - Translator
- Extensive database
- Word embeddings to deal with extra noisy data
.png)
.png)
.png)
We suggest you to create a separate virtual environment running Python 3 for this app, and install all of the required dependencies there. Run in Terminal/Command Prompt:
git clone https://github.com/purplepotion/Smart_Grievances_Redressal
python3 -m virtualenv venv
To run only the model:
* cd multi-class-text-classification-cnn/
* python data_helpers.py
* python text_CNN.py
* python train.py
* python predict.py
To run the dashboard: Python index.py
In UNIX system:
source venv/bin/activate
In Windows:
venv\Scripts\activate
To install all of the required packages to this environment, simply run:
pip install -r requirements.txt
and all of the required pip packages, will be installed, and the app will be able to run.
Run this app locally by:
python index.py
Open http://0.0.0.0:8050/ in your browser, you will see a live-updating dashboard.
Key words are displayed according to the department and the frequency graph of the complaints are dispalyed in the analytics dashboard