GCDH Summer School Digital Analysis with Digital Tools, Göttingen, 28th July to 1st August 2014
I have been lucky to participate in a great Summer School organised by the Göttingen Centre of Digital Humanities about Visual Analysis with Digital Tools. And I couldn’t have spent this week better! Even though one might think that temperatures around 30° C are tempting to exchange the classroom for a swim, I truly enjoyed all the hands-on sessions and lectures and our group evenings as well! In the following I would like to share some impressions.
The morning sessions were subdivided in two strands (Visual Network analysis with VennMaker and Gephi; 3D Documentation for Cultural Heritage), of which I followed the first strand. In this strand Michael Kronenwett and Martin Starck introduced network analysis and visualisation with VennMaker and Gephi. VennMaker is a great tool if you want to draw your own customized social network visualisations, whereas Gephi is a bit more flashy and it is easier to import data (but VennMaker people are working on a data import feature!). However Gephi, as trendy and popular as it is at the moment, is not updated anymore at the moment and it was on my Mac (with Mac OS 10.8.5) rather unstable. In the end I even lost the window where you can see the network visualisation… I guess only a complete new installation would solve this strange problem. Anyway, I have learned not only how to use these tools, but also how to interpret my nice looking shiny visualisations, and that some background knowledge of statistics always comes in handy…
Our collective meals after the morning sessions at the refectory of Göttingen University were followed by a divers array of afternoon sessions and evening activities.
Monday Dirk Rieke-Zapp of Breuckmann/Aicon introduced the whole group to the technical aspects of 3D Scanning with Structured Light. He had brought with him a 70.000 (!) Euro scanning installation and showed how it works in praxis. Every scanning situation is different and every excursion should be carefully planned. Test before you ship your equipment! A truly nice episode he told us, was about scanning the Mona Lisa. Do you know how many people it takes to scan the Mona Lisa? Answer: 5. The curator, the scientist, the scanning professional, someone to hold the tripod and someone to hold the cables…
The keynote on Monday evening was given by Daniel A. Keim of the University of Konstanz, who showed us some fascinating DH visualisation of the Bible, Mark Twain’s books, and Twitter emergency situation recognition… He drew our attention to the fact that data visualisation is especially difficult in the humanities as researchers in the humanities are often confronted with fuzzy data in which they try to see something new. He strongly encouraged DH researchers to combine the (dis)advantages of computers (fast!, accurate!, stupid!) with the (dis)advantages of humans (slow!, inaccurate!, brilliant!) and to stay adventurous, innovative and not to forget to have fun also.
On Tuesday afternoon, Norbert Winnige and Alexei Matveev of the MPI MMG introduced us to the research into visualisations at their institute, especially how to visualise migration “without arrows” which always remind a bit of battle maps. Alexei Matveev demonstrated in a hands-on-session how to make visualisations using Processing. This is quite tough stuff for people who are not used to programming, but he did a great job! He also uttered the wise words, that you should not be seduced by “tool trends” , but always consider carefully which tool will serve you best to do the job. (And if you haven’t found the right tool yet, have a look at the valuable list of DH-tools provided by Project DiRT.)
But we did not only dive into visualisations for the humanities and social sciences! On Wednesday afternoon Sven Truckenbrodt of Rizzoli Lab explained more about his lab’s last scientific breakthrough, the first 3D Visualisation of a synapse, which was even featured in Science! He gave an inspiring lecture on the history of visualisations in the sciences and what can go wrong and whether you can trust them at all. He strongly recommended the book Objektivität by Lorraine Daston and Peter Galison (there also seems to be an English translation). It is good that also in the sciences researchers start appreciating visualisations, because they say so much more than “a big chunk of data on a sheet”! But 3D visualisations are also very expensive. Their model comprises 5-6 years of pure research, an incredible amount of time of Burghard Rammner who programmed the visualisation, and a lot very expensive super computer computation time! You might want to watch the video in the website of Science!
After the spacy 3D synapse a team from the SUB (Stefan Funk, Stefan Funk and Ubbo Veentjer) took the stage to present the Dariah-DE Geobrowser. With the Geobrowser you can visualise time-space-relations for example in a network of letters very easily and very beautifully. You can upload different datasets and overlay and compare them if you, let’s take an example, want to compare the spread of the books of two authors. It works with the Getty Thesaurus of Geographic Names to recognise the place names and to put them on the map (which works even with historic place names), but you have to check carefully. Sometimes the first option in the Getty Thesaurus seems to be a quite US-centric, why else would it put Edinburgh initially in the States?!? The city tour in the evening was “boycotted” by a city run, so we were not really able to walk a lot outside, but definitely our visit to the old city hall was a highlight! My first time ever with a “medieval” key… And a key it was, I guess it was some 30 cm and 0,5 kg!!!
Thursday afternoon was completely dedicated to an introduction to R. The task to lead an already a bit tired crowd of DH apprentices to places where almost none of us had been before was boldly taken over by Andreas Cordes (Institute for Psychology, Göttingen). R and R Studio are a bit tricky to install, so this took some time. I have to admit, even though R had been developed as an easy programming language for non-programmers, it still requires you to truly think “computational” to get the job done. The good thing is, that the R community is very active and very willing to share little routines to get a specific job done (these are called packages ;-)). We all were really impressed of what R enables you to do, so learning the basics will really provide you with a powerful tool for all kinds of analyses. If you understand Dutch, I would strongly recommend reading for example Karina van Dalen-Oskam’s inaugural lecture or some of her other publications to which my attention was drown by Astrid Kulsdom. I am very curious to see what else you can do with R for Computational Literature Analysis.
The last day’s afternoon was dedicated to the presentations of the achievements of the two Summer School strands. The 3D visualisation strand had set up a virtual museum of their scans which was truly impressive. My strand, the network analysis strand, also showed some of our visualisations and Ingo Boerner and me even demonstrated live how you can make a very easy, but nevertheless quite impressive analysis of your Facebook network with Gephi. I had found a tutorial for this on Youtube, so if you like to try it out, just follow this link to the Video of Data J Lab of Tilburg University. As we also learned during this week about sensitive data and not intruding other people’s privacy (anonymisation!), I won’t post my result …
Many thanks to the organisation team of the Summer School (Frank, Ele, and Andrea) and all the presenters and participants who made this an unforgettable rich learning and fun week!