An interview with…Enrico Bertini

Today, before the Football World Cup Final, I thought was nice to listen to the voice of an Italian guy

His name is Enrico and he’s one of the few Italian researchers that works on infovis from a long time.

1 - When, how and why have you decided to research on infovis ?

I don’t remember exactly but I’m sure everything originated from this Howard Rheingold’s book “Virtual Reality”, which I read in 1994. It described research carried out by various people in computer science labs around the world. I was fascinated by the description of researchers struggling to find appropriate ways to represent reality. In that book I was deeply impressed by Prof. Frederick Brooks’ interview. He spoke about his idea of a computer scientist as a toolsmith: “a toolmaker succeeds as, and only as, the users of his tool succeed with his aid”, criticizing AI research and advocating for a more engineering-oriented and pragmatic approach to applied research (http://www.cs.unc.edu/~brooks/Toolsmith-CACM.pdf). The rest came out naturally: I liked computers, graphical interfaces and wanted to study the way people interact with them. Information Visualization was the only discipline that had a perfect mix of all these ingredients, technical and not. Then, when I was about to start my master thesis I was desperately searching for a professor interested in this topics. Finally, I discovered two professors in my department in Rome which had worked on a prototype called D.A.RE, an interesting prototype to visualize multi-dimensional data in a 3D scatterplot, and I asked for a thesis. I did my master thesis on visual data mining and then continued doing research on infovis for my phd, mainly on clutter reduction techniques.

2 - Which is, to you, the most interesting project you have worked on and why ?

I don’t know … I enjoyed all the projects I’ve been in. The current one is very interesting: we are developing novel visualizations to analyze users’ activity in private networks of large corporations; it’s a collaboration between NEXThink, a company in Lausanne, Switzerland (http://www.nexthink.com/) and the University of Fribourg, where I am now. There is an enormous log record of users and the applications they use, and the system is meant to support both administrators and business people to see what happens in the network and to study new security policies. It’s extremely challenging since data is very large and dynamic and it’s very hard to stay focused on few selected goals. It’s crazy how many different solutions and prototypes you can develop on top of the same data and tasks. The problem is to find the most useful ones. I think this is one of the major challenges of infovis: going beyond the beautifulness of pictures and build things that really make people successful in their own work. I also enjoyed many aspect of my phd thesis work, which was on clutter reduction. I liked the originality of the approach and the idea to understand what clutter is; a very ambiguous concept indeed. For example, I enjoyed conducting some experiments to understand how people perceive the number of dots in a computer screen (numerosity) and using these results to guide a visual optimization algorithm. I loved the necessary mix of knowledge required to conduct such research.

3 - On what themes are you working now ?

As I said, my main activity now is visualization for network security. On a more theoretical level, I’m also continuing my previous research. I’m working on a new classification of infovis techniques to describe clutter reduction in a more structured format. It’s an aspect of my phd thesis I wanted to develop further. I’m also intrigued by the idea of using visualization as a way to deal with personal information: emails, sms, documents, places, etc. There seems to be and interesting trend in this direction.

4 - If you should suggest an education programme to a student interested in InfoVis, which University Course do you suggest him/her ?

InfoVis is multi-disciplinary in nature, so there are many ways but it’s tough. Italy is the reality I know better, so for Italian guys who want to work on infovis the best is to get basic courses on human-computer interaction, design, and computer graphics (which is quite hard to find the whole three in the same faculty). As far as I know, the faculties where topics related to infovis are taught are: communication science, design, computer science/computer engineering, but none of them give the whole spectrum of knowledge. My suggestion is to be very strong in one basic discipline and develop the rest by reading relevant books and papers from conferences like: InfoVis, UIST, CHI, etc. In my opinion the best is to start with the intro chapter of the classic: Card, Shneiderman, MacKinlay’s “Readings in Information Visualization” and then read (many times!) all Tufte’s books and Colin Ware’s “Information Visualization”, which is my favorite.

5 - How do you think infovis researches will evolve in your country for next years ?

That’s a hard question since infovis is not very much developed in Italy. I see interesting trends from the area of graphic design which is historically strong in my country. There is a long tradition of very very good designers here. From the area of computer science, I see most people is still quite skeptical about infovis, or hci in more general terms. Researchers think it is not “scientific” enough, which is a strange way to look at the world today.

Thanks for the interview.