Tweet Sentiment Analysis

Developed a deep learning neural network for a tweets of US airlines customers. This dataset is avaialble on Kaggle.

https://www.kaggle.com/crowdflower/twitter-airline-sentiment

The LSTM Recurrent Neural Network based Flask Web App that classifies the sentiment of Customer Textual Review as Positive or Negative.

Firstly, developed a LTSM Model by using Flask Framework and then dockerized it. Later, streamlit app is created to show the gist of sentiments.

Getting Started

To get the app working locally:

1. Clone or download the repository locally.

2. Within the Tweet_Sentiment_Analysis directory, create a virtual Python environment with the Terminal command python -m venv sentiment where sentiment is the name of your environment. You can choose any name.

3. Activate the virtual environment with the command `

   flaskapp\scripts\activate.bat

4. Then run the command pip install -r requirements.txt (In case of error in Windows at this point, you need to set the LongPathsEnabled Registry value to 1. See here)

5. Next, set the FLASK_APP variable to app.py and FLASK_ENV to development by running the following command (for windows)

    set FLASK_APP=app.py

6. Also, set the FLASK_ENV to development by running the following command (for windows)

   set FLASK_ENV=development

7. And finally, run the command python -m flask run to start the app

8. The terminal will output the local web address and port where the app is running. As an example, this might be http://127.0.0.1:5000/. Now, open a web browser and go to that web address.

9. You can also access it typing by http://localhost:5000/

Docker image : https://hub.docker.com/r/tusharbedse/flaskapp

Docker Pull Command : docker pull tusharbedse/flaskapp

Prerequisites

You will need (Python=3.8)Python3 installed on your local machine.

Built With

• Python - Programming language
• Tensorflow - RNN Model
• Flask - Web Development Framework
• Pandas - Data Manipulation and Analysis

Interface Sample