Hacks built on Dandelion API during Hackference 2014: context matters

We’ve sponsored an interesting event, that took place in Birmingham during the last September: Hackference 2014.

Hackference 2014

Hackference is a three day technology driven event in the heart of Birmingham. The first day conference will include several talks from leading tech specialists. This will then be followed up by a two day hackathon.

The conference was very interesting, with some cool presentations, like the one titled Connecting to the pulse of the planet with Twitter APIs, or “Hacking through time and space“, that was very funny and insightful.

Not to mention the one on hacking dataviz, that cited R, and the one of Vagrant: two tools heavily used inside SpazioDati, to manage our data curation workflow.

Hackference Sponsors

So, who won the challenge?

We’ve sponsored a challenge on the best contextual experience built on Dandelion API. Because context matters: especially thinking about the overall user experience. That’s the point: we need to take care of the contextual experience idea.

You may hear the expression.. “If content is king then Context is queen”, No? Well, content it is still the king, but it is the context, which provides the framework to provide relevance to content.

And now less words and more facts: the winner is Melinda Seckington (@mseckington), with her hack titled “Tweet Tracks“!

“A few weeks back I attended Hackference 2014, where I came up with the idea of Tweet Tracks.
I love attending hack days and getting to play around with APIs I haven’t used before. So when I saw what was possible with the Dandelion API I thought it would be fun to see what I could do with it. My initial idea was to complement my Twitter home timeline with images of the objects that people were mentioning. It worked, but having just images wasn’t really that interesting.
I then realised: what if I combine it with Spotify? What if I could discover new tracks based on the tweets from my friends?
A Tweet Track example
So that’s what Tweet Tracks turned out to be: it uses Dandelion API to extract entities, like places, people, events, etc, from tweets on Twitter and then uses those entities as keywords to find tracks on Spotify. Initially it just did it based on my own home timeline, but I soon added a second tab where anybody could generate a playlist based on the tweets from any Twitter user.
It’s cool seeing the results of which tracks it finds: sometimes it’s really obvious, other times I’m curious as to why it found that track for that tweet. Plus I’ve discovered music I had no idea existed!” – Melinda Seckington

Give Twitter Track a try: is really fun!

Not only one winner, but two!

Last, but not least, we’ve decided to give a small prize to the team behind the TMTracker hack. They have used our Text Classification API without any documentation!

TmTracker example

TMTracker aims to allow you to monitor the public perception of your brand, or organisation, by periodically scanning news feeds, including Twitter and the Guardian, and scoring them based on the sentiment of the story or text. Using Dandelion API we were able to categorise the stories to provide greater insight into the data, allowing users to see public opinion based on related terms, or specific area of interest. Using this data, companies or organisations could determine which areas they were perceived to be doing badly at, so they could work to improve their reputation in that area, or what they are doing really well, so they know how they can play to their strengths.

For example, we were able to scan and track public opinion of Scotland, and then cross-reference that with the related term “Alex Salmond” to review the overall public perception of Scotland in relation to “Alex Salmond”. We could have further filtered these results to specific areas, such as Education, Environment or Health.
Integrating the Entity Extraction API was quick and painless, with a single HTTP request and parsing the returned JSON data to allow us to categorise the data.” – Stevan Goode

So, the prize is divided into two parts:

  • 400£ for Tweet Track
  • 100£ for TMTracker

Congrats to the winner teams, many thanks to Mike Elsmore, and to all the participants at hackference: it was a really, impressive event!

Related sources