CoviVax Sentiment Analyser

• 

  Correct classification along with other headings

• 

  Correct classification along with other headings

• 

  App Outro

• 

  Negative review entered

• 

  Positive review entered

Inspiration: Our inspiration stems from the idea of saving resources for our country during the unprecedented times of covid-19.

What it does: It pains us to know that more than a 1000 doses of vaccine are wasted due to the preferential issues of the general public, which is why our app helps hospitals to understand and review the preferences of an individual in a qualitative manner to optimize their use of the vaccine batches, saving the resources for the needful.

How we built it: We built this application using a Kaggle dataset, which was based on the vaccine sentiments of people extracted from twitter tweets. We analyzed the dataset using libraries such as numpy, pandas etc. For the NLP processing, we made use of the tfidf vectorizer and created the model via logistic regression. In the end, we created the UI using a library called PyWebIO, which helped us to create the UI efficiently.

Challenges we ran into: Due to the low computational powers of our computers, we had to think of an alternative route for text processing, and during the creation of our UI, we ran into a lot of system errors and exceptions, which were ultimately solved by our team using out-of-the-box ideas and tricks.

Accomplishments that we're proud of: Working with such a large project, like this, for the first time, our team is especially proud of the way we handled all the errors popping up in our project, and also, learning the new PyWebIO library was especially challenging as none of us were accustomed to the working and syntax of the library. Overall, we are proud of what we've given birth to.

What we learned: We learnt about the new PyWebIO library for generating UIs for our ML model, and we gained deeper insights into the tfidf vectorizer. We were also lucky to learn about various types of exceptions and how to handle them.

What's next for CoviVax Sentiment Analyser: Our app requires a cloud backup, so that any review given by any user can be stored, and hospitals/clinics can use the data to optimize the use of vaccine batches.

Built With

• github
• logisticregresion
• numpy
• pandas
• python
• pywebio
• scikit-learn