Inspiration

Since , I am obsessed with Twitter , so I thought to make a dashboard which will do analysis of the sentiment of the tweets.

What it does

It basicallly analyses the polarity of the tweets as : most positive, most negative and neutral.

How we built it

I built by using Logistic Regression and checked its accuracy of my machine learning model. Furthermore, I cleaned the dataset and removed stopwords and done its preprocessing. Initially, I extracted the tweets by using a module snscrape.

Challenges we ran into

Cleaning of tweets and accuracy betterment were the challenges we ran into overall.

Accomplishments that we're proud of

Finally , we succeeded in extracting the sentiment of the tweets.

What we learned

How to clean a dataset and preprocess it and how to increase the acccuracy of machine learning model by using different ml algorithms which fit better.

What's next for Sentilyzer

I am looking forward to make it language specific too so that it can detect the languages and extract sentiments from it.

Also emoticon analysis will be a cherry on the top of the cake. Looking to implement it soon.

Built With

  • logistic
  • matplotlib
  • nltk
  • pickle
  • plotly
  • python
  • regression
  • seaborn
  • sklearn
  • snscrape
  • vadersentiemnt
Share this project:

Updates