NLP-YouTube-Comments-Sentiment-Analysis

Build + Deploy a Sentiment Analysis Model(on Heroku using flask) to classify YouTube comments for brand image improvement

An end-to-end toolkit on building a sentiment prediction model with a Jupyer notebook and deploying model on Heroku using flask. Our use case here is review classification of YouTube comments on Samsung Galaxy S23 Ultra into positive, neutral & negative and suggest moddifications to the product to improve brand image of Samsung. I have scraped comments using youtube_comment_scraper_python library by bot-studio.

Deployed model on Heroku

Sentiment Analysis ISI Chennai 2023

I suggest the following modifications to ”Sumsung Galaxy S23 Ultra” developing team based on the negative comments recieved

• To include expandable memory card slot for upcoming version. • To use the better processor for upcoming version. • To include more features than previous model for upcoming model.

How the model works!

Step by step Implementation:

• Scraping comments on Samsung Galaxy S23 Ultra's official film available on YouTube using Python • Labelling data as positive or negative based on polarity scores • Transforming data: Using lemmatization module; converting alphabets to lower case; removing new lines, punctuations, special characters, stop words; tokenization • Encoding: Labels to numerical form; comments to vectors using TfidfVectorizer • Balancing the dataset • Machine Learning model: Splitting the data into train and test set, feeding the data to Support Vector Classifier • Model pipeline: Creating a pipeline which passes the data first into TfidfVectorizer and then to SVClassifier • Model evaluation using accuracy score • Saving the model pipeline as pickle file • Making index.html and flask app(app.py) • Committing the files to Github • Creating an app on Heroku and connecting it to Github repository • Model deployment on Heroku

Steps to run on Windows for local Deployment

• Prerequisites: Python 3.9 (ensure Python is added to PATH) + Git Client

• Open GIT CMD >> navigate to working directory >> Clone this Github Repo

  git clone https://github.com/abdulwahed98/NLP-YouTube-Comments-Sentiment-Analysis.git  
  

• Open Windows Powershell >> navigate to new working directory (cloned repo folder)

• Create a virtual environment

  • install virtual environment

    pip install virtualenv
    

  • create virtual environment by the name ENV

    virtualenv ENV
    

  • activate ENV

    .\ENV\Scripts\activate
    

• Install project dependencies

  pip install -r .\requirements.txt
  

• Run the project

  python app.py
  

• Look for the local host address on Powershell screen, something like: 127.0.0.1:5000 >> Type it on your Web Browser >> Project shall load

• Try out your Amazon Alexa test reviews and look for results

• To close >> Go back to Powershell & type ctrl+c >> Deactivate Virtual Environment ENV

  deactivate
  

Steps to run on Mac for local Deployment

• Prerequisites: Python 3.9

• Open Terminal >> navigate to working directory >> Clone this Github Repo

  git clone https://github.com/abdulwahed98/NLP-YouTube-Comments-Sentiment-Analysis.git  
  

• Navigate to new working directory (cloned repo folder)

• Create a virtual environment

  • install virtual environment

    pip install virtualenv
    

  • create virtual environment by the name ENV

    virtualenv ENV  
    

  • activate ENV

    source ENV/bin/activate
    

• Install project dependencies

  pip install -r requirements.txt  
  

• Run the project

  python app.py
  

• Look for the local host address on Terminal screen, something like: 127.0.0.1:5000 >> Type it on your Web Browser >> Project shall load

• Try out your Amazon Alexa test reviews and look for results

• To close >> Go back to Terminal & type ctrl+c >> Deactivate Virtual Environment ENV

  deactivate