Inspiration

This is part of the ConUHacks YellowPages challenge

What it does

It streams tweets in real time and responds to #askYP

How we built it

We used tweepy to stream the data from twitter and filter for #askYP. Then using google-cloud natural language processing API we managed to extract consumer goods, locations and organizations. Lastly we forwarded the keyword provided by the natural language processor to the yellow pages API to retreive places relevant to the user's request.

Challenges we ran into

Getting the authorization tokens to work took us some time as the twitter documentation was outdated and it took a lot of research to get an access to the API. Using tweepy was pretty straight forward. Google-cloud was somewhat challenging because it required a lot of installs and configurations for the API to work. Once we had the authorization working using the natural language processor was super easy and intuitive. Lastly, we started off on the wrong foot with the yellow pages API trying to get a key for a merchants app. We realized pretty late that we had missed an email from the team that provided other means of using the API at was a lot more convenient and didn't requite authorization

Accomplishments that we're proud of

One team member didn't know any python going into the hackathon and she learned a lot in a short period of time. We also didn't have any experience with any of the tools/API used in this hackathon and it's always fun to try out new stuff.

What we learned

This was out first hackathon and we're really happy with it. We now have a pretty good understanding of some features of the twitter API. We also realized how easy it is to use powerful tools provided by google like the natural language processor.

What's next for twitterBot

Adding localisation will greatly improve the quality of the suggestions provided by the bot. Unfortunately we started using the correct yellow pages API a lit bit late and didn't have the time to get a good hang of it.

Built With

Share this project:

Updates