What's better than distributed computing? Low-powered distributed computing!

What it does

A cluster of pis doing distributed debate analysis. A cluster of pis are wrangled together and run Hadoop in order to process presidential debate data.

How we built it

We strapped some Raspberry Pis together and set them up as Hadoop nodes. From our head node, we loaded and processed presidential debate data from the first two Republican debates and had the cluster process the text and crunch some numbers. From there, we developed a Django app to display the data in D3!

Challenges we ran into

First time Pi & Hadoop development - the Pis were not always cooperative and getting Hadoop up and running was a little rough at first.

Accomplishments that we're proud of

We have a working cluster! We found some interesting data in the debate transcripts.

What I learned

Lots of Pi and Hadoop knowledge

What's next for Elephant Pi

What's not next for Elephant Pi??? (Pushing it to prod, actual adoption, speed)

Share this project: