Have you ever wondered about the progress within hackathons? How do those projects progress,
what do they achieve, how many teams are after the grand prize?
And, what happens after the winners are announced, where the winning teams end up, and
then of course what happens to the results? In short, is there a story to tell about
challenges?
I found a while ago a list of competitions run by Kaggle - a start-up of scientists
that specifically run data-science competitions - but I could not figure out (just looking at a longish table) what the most popular scientific topics are or whether there is a correlation between the topic, the number of participating teams, the prizes and the
length of a challenge, etc.
Thus, I decided to give a try and make an interactive visualization using the data available. I also wanted to check out what the new d3.js framework is capable of, so consider this work more like an experimantation in the sandbox.
The design metaphor for the visualization is a "progress clock" that shows individual projects, in this case competitions, and how they progress towards
the finish. All competitions start in the middle of the clock and progress towards 12 o'clock. The distance from the center (the radius) is
proportional to the absolute time spent on the project so far. Consequently, the distance from the edge of the circle shows the remaining time.
The position around the edge of the circle (the angle from the start, basically the "hours") is the relative time spent on the project so far.
Three o'clock is 25%, 6 o'clock is 50% of the time given for a challenge, etc.
The circles represent discrete moments in time and the size of a circle is proportional to the number of participating teams. Since the Kaggle
table presents only the number of teams for the current date, I had to invent some numbers for the participating teams between the start of a
project and the current date (surely they have this data, I just used some heuristics to describe how the number of teams grew since the start
of a competition). The colors encode the "sciences" or topics the challenges belong to, an arbitrary classification that I will revise at some point.
Also, there are only six competitions in the clock at this time, will add some functionality to extract the data automatically from the table on Kaggle's web page.
In so doing, hopefully an engaging interactive visualization is presented for the user, and more importantly, a story or two become visible. Enjoy.
Péter