We've always been loyal users of TripAdvisor, Expedia, Yelp among others when we go traveling to place we have no idea about. Although this may not always be a bad thing when it comes to places like Greece and Rome, the recommendation seemed to need a huge upgrade!
What it does
We're changing the way it recommends places to visit. Using the Personality Index API from IBM BlueMix, we analyse your tweets (default) and figure out your personality to work on what would best suit your kind of travel. Doing a layered Sentiment Analysis, we worked our way from scraping our (formerly) favorite sites for reviews, and correlating the Personaliy Index of the Reviewer as well as the Sentiment of the Review. Based on the correlation, we then recommend a list of places that you would love to be at. Last but not least, we used the Amazon Alexa to be our Travel Buddy giving us cool use cases for a kick-ass conversation.
How we built it
A Pinch of IBM-Watson and a Dash of Amazon AWS & Lambda.
Challenges we ran into
Finding reviews was the hardest. Due to all the website's business models, none of the data was readily available to take from the APIs
We added a #HackRice Mode that enabled us to make "cooler" suggestions that would cater to the student crowd. :D
What's next for Wayfarer
We could only do so much of analysis over the weekend, so to make this the absolute Travel guide, we're hoping to implement more effective Machine Learning models to make better recommendations at higher accuracy.
And who knows, the places recommended might even scare you.