Inspiration
A few weeks back, on a boring Saturday afternoon, I was sitting on my couch, drinking a cup of coffee and wanted to do something. What, I did not know but something that would be fun, adventurous. I got my laptop out to search for this 'something' and it was a really bad experience. First, I did not know what to Google, as in 'Ok Google show me something adventurous around me' makes no sense to it. And secondly, when I was finally able to search for some specific events it threw 10 links at me which just made me go back to my couch. That is how the idea, of having a chatbot to specifically show you things happening around you by understanding your feelings and personality, came through.
What it does
The messenger chatbot understands the feelings of the user i.e. creative, leisurely, adventurous etc. and suggests relevant activities, workshops, meet ups in his locality. It can also help out with specific dates and budget.
How we built it
We trained a named entity recognizer using wit.ai and the rest of the bot was built using the messenger api in node-js. For maintaining a database of current events we used mongodb, which was populated by scraping websites (which legally allow themselves to be scraped) using python.
Challenges we ran into
One of the major challenges was to make the nlp system understand something as abstract as feelings. For solving this problem we painstakingly taught it hundreds of sentences and finally got satisfying results.
Accomplishments that we're proud of
Being new to this field and starting from scratch, getting a chatbot up and running in a span of 20 days is something we are really proud at having accomplished.
What we learned
We learned a lot ranging from basics of natural language processing, web scraping to using various cloud based services to create an end to end application.
What's next for Find things to do near you
We plan to tie up with people having top trending facebook pages in this domain like 'Thing to do in Mumbai' etc
Log in or sign up for Devpost to join the conversation.