The ACDH virtual hackathon series: Open Data for Open Source solutions

At the beginning of 2019, the Austrian Centre for Digital Humanities of the Austrian Academy of Sciences (ACDH) started an experiment: For the first time ever, we published the calls for participation in three virtual hackathons. The hackathons would focus on Open Data sets publicly available online, and the tasks to perform on these data would involve creating Open Source code. Each of the hackathons had a special theme and was co-timed with events that involve an aspect of Openness. These events also inspired the choice of the respective data sets. The best contributions would be determined by an international board of judges and receive cash prizes. The criteria for judgement were the following:

• Creativity, innovation (e.g. Is the approach/idea new and unique? Does it do something that hasn’t been done before? Does it provide new insights into the data? Does the hack provide a new/faster/clearer solution to the old problem?)
• Accessibility, reusability, reproducibility (e.g. Is the code properly documented? Is the technical approach reproducible?)
• Elegance (e.g. Is the code easy to modify and reuse? Is it readable for others? Is modularity considered in the design? Is the code simple and concise?)

The hackathon series was funded by the Austrian manifestations of DARIAH and CLARIN, CLARIAH-AT. The initiators and project leaders were Tanja Wissik and Vanessa Hannesschläger.

Not only in the context of Digital Humanities but also in other research areas the Open Data movement is gaining momentum. One example of this development is the International Open Data Day which is becoming more prominent every year. The City of Vienna has launched its own annual Open Data happening. But also in the context of DH Open Data is becoming a vital part of the research process as projects such as ELEXIS are opening up (research) data. One of the benefits of Open Data is that it can be used by anyone for any purpose which makes it an ideal resource for hackathons.
Usually hackathons take place on site, participants are given tasks to be solved within a given timeframe in a fixed location. This requires the programmers to be flexible and available and to have access to travel funding. A virtual hackathon on the other hand offers people all over the globe the possibility to participate and contribute without having to travel.

ACDH hackathon series call for contributions

hackathon #1: the ELEXIS hack

The first hackathon was carried out in cooperation with the European Lexicographical Infrastructure (ELEXIS) and focused on lexicographical data. In this hack, the task was to develop a creative mode of processing it, e.g. by enriching it, visualizing it, doing statistical analysis, or integrating it with other resources (e.g. LOD). The submissions were to be handed in by the end of the first ELEXIS observer event which took place at the ACDH in Vienna from 18 to 20 February 2019. The dataset to be worked on in this task was the Digital Dictionary of Tunis Arabic.

In response to this first call, Thomas Burch and Li Sheng created TAGEF-Net, the “Tunis Arabic German English French Network”. They implemented a networks visualization application based on the given TuniCo dictionary to show the relationships between Tunis Arabic words and their translations in German, English and French. Irina Treichl on the other hand chose a different approach. She developed an interface for filtering and visualising the data by radical consonants, their position in the word and the roots of words in the dictionary, TuniCo Roots.

The judges (Peter Andorfer, Saranya Balasubramanian, Thierry Declerck, John McCrae, Dieter van Uytvanck) of this hackathon were equally impressed by both these approaches, as they expressed in their ruling:

TAGEF-Net serves as a general purpose exploratory tool and TuniCo Roots allows more specific ways to explore patterns in the lexicon using consonants and roots. Both submissions have different strengths and weaknesses and owing to the difference in the ideas themselves, we are unable to pick a winner and runner up. Therefore we recommend both submissions to share the first place.

hackathon #2: the Open Data Day Vienna hack

For the second hackathon, the ACDH cooperated with the City of Vienna and chose a set from the city’s Open Government Data platform to be processed in this task. The task was to be completed on the Open Data Day Vienna 2019 (28 February 2019). The aim was to develop a creative mode of processing cartographical data showing war damage to buildings in Vienna compiled by the City of Vienna in 1946.

This task inspired Asil Çetin to develop vienna1946, a front-end focused solution including a demo app that allows to explore the bomb damage to the City of Vienna during World War II on a 1946/2019 time travel map. Kevin Stadler and Ana Jeličić on the other hand focused on extraction of information from the map in their contribution Bomb damage. Their contribution consisted of a method for data cleaning and online map tile generation and an interactive web map to explore the historical data.

The judges (Mateusz Żółtak and Ksenia Zaytseva) found these two submissions to complement one another and came to the following ruling:

The nice thing is that both submissions can be seen as complementary: The better dataset being an outcome of Kevin’s work can be easily incorporated into Asil’s solution (only one additional line of code is required). Unfortunately it also makes both submissions incomparable with each other. Taking into account that both are of similar quality (details below) and we liked both of them, we apply for granting both submissions an equal first place.

hackathon #3: the International Open Data Day hack

The third and final hackathon of the series offered a task to be completed on the International Open Data Day 2019 (2 March 2019). For this final highlight, a choice of two Open Data sets was offered to the participants. The first data set to be worked on in this task was a collection of XML/TEI transcriptions of early German travel guides on non-European countries (Baedeker travel guides). The texts are tokenized, labelled with lemmas as well as part-of-speech-tags, and semantically annotated. The second data set was provided in cooperation with Europeana. It was composed of high definition images and full-text of 8 newspapers in German contributed by the Hamburg State and University Library Carl von Ossietzky to the Europeana’s project on newspapers.

Marta Palandri and Raphael Mitsch processed the Europeana newspaper data set and submitted The Quick War?, which aims to investigate how the sentiments of German newspaper articles that report on World War I changed in the course of the war. Martina Trognitz and Lukas Gehrig on the other hand worked on the Baedeker travel guide texts and created xmlTEIontheMap, which aims to visualize locations that occur in the Baedeker guide “Das Mittelmeer” (The mediterranean sea) on a map.

Despite the differences between these two very creative submissions, the judges (Julia Neidhardt, Gerlinde Schneider, Mete Sertkan) came to the following ruling:

“Since the objectives of both submissions as well as the data the build on are completely different, the submissions are incomparable to each other. In general, we liked both of them. However, taking into account the creativity, reproducibility, reusability and elegance of each of the submissions (see below), we apply for granting xmlTEIontheMap the first place and The Quick War? the second place.”

The ACDH team was thrilled to see so many diverse and creative submissions contributed to the virtual hackathon series. As all submissions are based on Open Data and were created as Open Source code, all work done in the course of the hackathon is openly and freely available to all. We hope to see further development inspired by these submissions and look forward to future hackathons.