Twitter-Summary-mini-Project

This project is to identify top trends in twitter and generates the summary of those trending hashtags in about 50 to 160 words using Text-to-Text transformer model.

Requirments:

• Python packages mentioned in Requirments.txt To install packages use

      !pip install <package name>
  

• Pre-trainined transformer model to be downloaded
• Twitter credentials like API Key, Access Tokens,etc Steps for getting Twitter credentials:
  • Create a twitter account incase if you do not have already
  • Go to https://apps.twitter.com/
  • Click 'Create an app', this will help to create a developer account
  • Once you fill all the mandatory details you will be provided with required keys, please save those keys locally in a notepad for later use

Flow of the Project run:

Step 1:

Identify the top trends in twitter for the current day with respect to the user defined location.

Step 2:

Select the top n number of trends in the user defined language based on the volume of tweets available at the point of extraction.

Step 3:

Extract the tweet data for those selected trends

Step 4:

Do bit of data cleaning & processing to cater well for the model

Step 5:

Put all the individual tweets together as a string chunk

Step 6:

By leveraging pre- trained model’s ability, generate text summary of the tweets in about 50 to 160 words along with Word cloud graph for better understanding of the context.

For more information on the NLP model: Refer : [Model Blog] (https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html)

Future Scope:

Able to work with multi language tweets using google translate API, comprensive web app for the ease of intraction, model tuning better for results.

Linkedin : Balaji Sri Raj.