Data visualization challenge at the first SatRdays conference
The conference was quite a success, with 25 international speakers and a soldout attendance of 180. As part of the programme, we ran a data visualization challenge, and here I’ll summarize the results and also share some thoughts on dataviz and R.
The dataviz challenge
The challenge rules were quite simple: we published a dataset consisting of 18K+ entries on the flights of the Budapest Liszt Ferenc airport, sourced from the Hungarian Central Statistic Office. The participants were asked to submit a data visualization created from the data using only R.
While such contests are always engaging and fun, this time we also had two additional goals:
- First, we wanted to see how similar the methods and packages used by entries would be. Is there a common way of doing data visualizations in R? Or will we see a number of different approaches?
- We have already ran a very similar contest on the same dataset at our Budapest BI Forum conference a few years ago, where all kind of commercial data visualization software was allowed and used. We hoped that comparing the entries of the two contests would show how far the dataviz capabilities of R progressed.
I’ll get back to these points later, but let’s see the entries and results first. We have received 8 submissions in total, and all contestants were allowed to present her/his solution in 3 minutes at the last session of the conference. The audience then voted on the best submission.
All presentations were well received by the audience, and after counting the votes, the following 3 entries emerged as the most liked:
The 3rd place went to Thomas Levine, who created an animated video with custom-made music decoding the data points.
The second prize was won by Ágnes Salánki, whose visualization was aimed to those who don’t like the crowds (‘a recommedation system to introverts’).
And the first prize went to László Gönczy and his teammates at Quanopt Ltd. (János Oláh, Nóra Lengyel, Flórián Deé and Imre Kocsis) who created a beautiful dashboard showing different stories hiding in the dataset.
The most popular technology was Shiny by far, used in 5 of the 8 entries. Other R packages used includes plotly, ggplot, ggmap, ffmpeg and grid.
I’d like also to mention the effort by Romain Francois, who not just created a great visualization together with Cecile Sauder, but also published a revised and updated version of the dataset on github.
List of all entries (in presentation order):
Ágoston Török és Fanni Kling: Results, code
Tibor Szabó: Results, code
Tamás Markó: Results, code
Thomas Levine: Results (.mkv), code, summary
Zsuzsanna Szabó: Results, code
László Gönczy et al.: Results, code
Ágnes Salánki: Results, code
Romain Francois and Cecile Sauder: Results, code
There’s more to come
If you also like R or data visualization and enjoy visiting Budapest, check out the Call for Papers for the upcoming Budapest BI Forum conference (October 25–27). We will run the next dataviz challenge there, with a brand new dataset. The CFP closes in a week.