Its the holiday season and many of us want to take a break, get away and relax. Use safespot to find out how safe is your choice of outing during the pandemic. #besafe #vacation #relaxing #covid
It's the holiday season and people want a break from the stressful pandemic and staying at home regulations. Further data shows that COVID cases are exploding during the holiday season. The problem of helping people find safer spots to relax, travel reducing their risk of causing and contracting Covid inspired safespot.
Safespot asks user to enter the location they want to go. This can be a restaurant, vacation spot or even just a walk down the park. it then provides a safety score on a scale of 0-5. The safety score is calculated based on number of cases in the location city, vaccine sentiment in the state, safety measures followed a the location for instance masks are compulsory and social distancing guidelines and finally reviews scraped about the place from yelp and google. The user can use the information to decide wether the place is safe to vacation amid the Covid circumstances.
BACKEND
Vaccine sentiment: We used python tweepy to scrape tweets related to Covid vaccine at state. We then used flair and text blob libraries to perform sentiment analysis on the tweets. Used Sqlite to store this information.
Google and yelp Reviews: we used https://wextractor.com/ to scarped reviews of particular location and performed sentiment analysis on these reviews. For a easier read we summarized the reviews in the last month using torch and T5 model for abstractive summaries.
Number of Covid Cases and Death at city: https://www.postman.com/explore/collection/6337/covidti-api
Used Googlemap Api: to get location coordinates
FRONTEND
Used react to develop the web framework which takes user entreated location as input and display information related to asses Covid safety for that area.
Flask was used to coordinate between frontend and backend.
We found the development of the web framework challenging using react and Flask having little experience in the area. We struggled getting our backend and frontend working together. Moreover implementing web scarping and and coming up with features to calculate our safety score was also challenging.
learning more NLP and frontend stuff
x
x
x