English„I am going to participate in Data Hackathon in London.”
“OMG, are you a hacker?!” was the common answer I got after I told someone that me and Torsten were attending London Data Festival in June 24-26, 2016.
Data festival organized by Keboola (Petr Šimeček, Pavel Doležal, Martin Lepka) at Wayra UK started with a “business conference” part. Why the quotes here? Because the whole weekend was held in an extremely informal atmosphere so the conference was more like “ok, let’s sit together around one table and talk.“ The discussions were triggered by guys with BI experience in retail (Paul Cook, Martin Zeman) and hospitality (Matt Lukowski).
Afterwards, Data Hackathon part was kicked off. The organizers had ready interesting data sets from Erste and CCS. But since we were starting a day after the UK’s EU referendum, they decided to change the direction and focus on brexit data. From Friday evening till Sunday afternoon, we were crunching various data – results of referendum, a million of tweets with associated hashtags and sentiment/entity analysis done above them, crime statistics, demographics and other data sets.
Data was one source of fun, the another were the technologies – we played with Tableau, GoodData, Snowflake, Geneea, Keboola Connection and its API such as Geocoding Augmentation.
An interested group of people attended – from BI managers to startup new joiners, including one Canadian guy on his Eurotrip.
In parallel, workshop for data girls was taking place. Well, there is some disadvantage if you organize an event for girls only (= guess how many girls there were on our “mixed” event).
Let’s show two examples of our findings to give you a better idea of what we were doing:
1) It seems that people tweeting from iPhones and iPads were proportionally more for Remain then people tweeting from Web Twitter or Android app:
2) Or – as Torsten found out – “Leavers“ were in fact in their tweets most of the time using more negative language and “remainers” more positive language. Only when they won, “leavers” switched to positive. Also there was a negative peak when Jo Cox was killed (on both sides):
Some pictures from the event:
Data Festival kick-off by Pavel Doležal (author: Petr Šimeček)
Petr Šimeček (author of the picture), me, Jirka from Geneea and Torsten (from left to right)
Simplity selfie from the very end
Key takeaway? During open table discussions it was repeated several times from many sites and data hackathon showed it us as well: Data governance practices are important not only for traditional BI/DWH projects, but also for more agile data projects. Understanding and governing the data is a crucial element for success of all data intensive projects.
Ondřej Kubera
Analyst
[email protected]