We were really hyped about playing with the Amazon Echo, trying to get Alexa to say all kinds of crazy things. But even with all the functionality it does have, there are a number of things that it should have; this is where it all started.

What it does

Alexa hears her name, "tweet", and your message. We extract the message to send and get various data from it; this includes getting popular hashtags and associations with what you wrote. In addition, we used IBM Watson to get the emotions associated with the tweet. A response is sent back to Alexa who reads a message based on the overall mood of the tweet.

How we built it

We made the Echo side using Javascript. Using Twilio to relay the message to itself, we were able to get the message available to extract data from. To extract data, we made a Python text analyzer that extracts certain words from the string; afterwards, IBM Watson evaluates the emotions expressed. Once evaluated, the string is tweeted and a response is sent back to Alexa to end it.

Challenges we ran into

We had various troubles figuring out how to make what we needed. I had troubles learning Lambda, especially for Python, since there aren't very many Python examples available. I had to work with Javascript, which is a language I have very little experiences with. Rai had troubles coming up with the best approach to automatically generating the best hashtags for the post. We found various libraries that helped us figure out how to do it.

Accomplishments that we're proud of

We're proud that we worked on things that we've never done before or had little experiences with.

What we learned

We learned how to work with some new APIs and technologies such as Twilio.

What's next for Alexa Smart Tweeter

We're hoping to patch things up to work perfectly and to deploy it soon. We hope that people who own an Echo and love to use Twitter would try out our skill.

Share this project: