News

Datathon 2019 – The Story

On April 16th, the fifth edition of Datathon 2019 – the International Data Science Hackathon finished.

More than 140 participants participated in the two-day event, held in four locations in the world – Sofia, Bangalore (India), Jaipur (India), Beirut (Lebanon) and online. The task was to work with real data to solve cases of business and social significance. Six teams reached the Finals, two of them from Lebanon and four from Bulgaria.

The Cases & its Solutions

The first case – Sofia Air 2.0, developed jointly by Telelink, Denkstatt, Green Sofia, Sofia Vision, and the Sofia Development Association, focuses on identifying the different factors of air pollution in Sofia and the interrelations between pollutants that affect the overall picture of pollution in different locations in the capital. The case study is a continuation of a previous Datathon in 2018 when a model for a 24-hour forecast of air pollution was developed. The aim of the task at the Datathon 2019 is to achieve a sophisticated approach to exploring the different dependencies, correlations, and factors that determine air pollution. This is the next step towards an overall view of the social theme of air quality in Sofia, based on an analysis of different in-kind data.

➡ Sofia Air 2.0 Case 

➡ Sofia Air 2.0 Leaderboard

➡ Sofia Air 2.0 Solutions

The second case – Kaufland AirMap is related to the processing of images of products made in a shop, the task is to identify the items, their availability and location on the shelves. This would automatically inform staff of irregularities or relocation of products, as well as reduce the availability on the shelves of some of them.

Kaufland AirMap Case

Kaufland AirMap Leaderboard

Kaufland AirMap Solutions

The Finals

The finalists shared the pool of prizes including tickets for Europe’s innovation and tech event of 2019 – Webit , provided by the Municipality of Sofia (ARC),  access to training courses for all finalists on the Data Science educational platform 365 Data Science, a participation in the World’s Largest Developers Congress in Berlin by WeAreDevelopers and $ 1,200 AWS vouchers.

The third place went to Phoenix team (Dimitar Vasilev, Vichko Cholakov, Atanas and Peter Stoyanov), working on the photo object recognition case, the second winning team was the Wonder Gang (Ana Antonova, Kamelia Kosekova, Ana Popova, Izabella Taskova), composed entirely of ladies with an interest in Data Science, who worked on the case of identifying the air pollution factors. And the winner was the Chameleons team (Ivan Vrategov, Kalina Angelova, Krasimimr Ivanov, Stanislav Dimov), who also developed a model for identifying factors influencing air pollution.

Among the Finalists were also two teams from Beirut AI – team Predikt (Karim Eid, Joe, Marcel Affi and Abdallah Atieh) and team Internet of Kings (Jack Abi Aad, Samer Salameh and Rida Wazneh). Three additional awards were given – an award for the audience, the most active participant on Social Media and the special prize was won by the youngest participants(Evgeni Dimov, Kalin Doichev, Kostadin Kostadinov and Aneta Tsvetkova) from Technological School Electronic Systems, TUES.

Check out the Finalists’ Presentation!

Open Community

All models and results are shared on the Data Science Society website – an international community of people with an interest in and experience in Data Science, which has been applying the expertise of its members for various social and business purposes for four years. Among the most significant achievements is the development of models to identify fake and manipulative news as well as propaganda.

If you are part of a community, company or else and want to experience a Data Scinece Challenge as a host or as a case provider, we are open to ideas and suggestions – Contact us or write us directly at [email protected]

Share this

One thought on “Datathon 2019 – The Story

Leave a Reply