UNDERMINER

A hobby project built on top of Marvel's API, displaying a random comic book and 27 recommendations with the help of a custom designed data analyzer algorithm. Users can also search for a desired comic series / issue from Marvel's database, or suggest characters not associated with the issue on display. This version of the site is under development, the live, responsive (well, almost) version can be found at underminer.net.

Otthonterkep is a real estate startup I'd produced for my former employer, Central European Media & Publishing (CEMP). Besides starting the company, recruiting team members and supervising production, I made sure to bring code to the project personally: I've designed the basemap's tileset using OpenStreetMap data, and built all map-based visualisation prototypes for the final product.

My submission with Daniel Molnar for a data visualisation contest by Boston's Hubway bike sharing system. I'd designed a Tilemill / Mapbox tileset based on a number of data sources, including mapdata from Oliver, MassGIS's awesome online data repository, and satellite images for shaded relief from NASA’s Shuttle Radar Topography Mission. Our submission never made it to the finals, it was a fun project though!

One example of the interactive data visualizations I'd built for index.hu, Hungary's most visited general interest portal - and my former employer. This viz shows data published by the National Tax Agency: all companies and citizens are on display who were late (as of June, 2013) with their taxes over a period of 6 months and a sum of 40.000 USD (citizens) or 400.000 USD (companies). Unfortunately, lots of data were available.

This video is about how the number of tobacco selling points had changed over a course of a couple months in Hungary, after the government nationalized the tobacco-selling market. Besides showing the capital, Budapest in details, I also displayed the 19 counties' chief-towns of Hungary and some larger regions. Fun fact: Tommy Ramone was also born in Budapest. The video was shot for index.hu.

Also a teamwork with Daniel Molnar, this time for Information is Beautiful's Hollywood Visualisations contest. We scraped some contextual data and merged them with the sets provided for the challange, so were able to examine box office numbers / country, and made up some exciting metrics (e.g. national gross / 1000 capita). Trailers are also included! The viz made it to the top five in the contest.

A paper based on the project I did for origo.hu, where I'd worked for 7 years in various positions, mostly in the editorial. The R&D coproduction with University of Szeged's Artificial Intelligence Research Group was successfully tagging the newsportal's archives consisting of 400.000 articles. The paper was published at the 19th European Conference on Artificial Intelligence (ECAI 2010), where it was shortlisted for best paper.