Big Data, Analytics, Networks, and Predictive modeling

The latest edition of Information System Research has an editorial by Ritu Agarwal and Vasant Dhar, who discuss IS research in the emerging areas Big Data, Analytics, Data Science, Network Science, and Predictive Analytics.

Data science, randomized controlled trials and networks are rapidly changing what we can study and how we can study it.

This is yet another wake-up call to the IS community. But who is listening?

I fully agree with what is being said in this piece, and I urge you to read it carefully.

6 degrees of separation in the boardgaming industry

In my introductory course on network science I usually use the popular "Hollywood actor network" to introduce students to the topological characteristics of social networks. This network is the basis for the Kevin Bacon-game, where you try to get from any actor to Kevin Bacon, by movies that they have appeared in together.

The Hollywood network is great (and a classic), but I prefer a game night over a movie night. Also, I like the idea of using networks to describe relationships between products in an industry. Long story short, I have constructed a Boardgame Network where boardgames are connected by the people who create them (artists and designers). This site is basically for my students, but others might find the idea interesting as well.

You can explore the Boardgame network here.

A PDF with statistics and images is here.

German soccer: A network summary

This network shows the results of every soccer match played in the German Erste Bundesliga from 1963-2014. Click to enlarge. A high-resolution PDF is also available.

The network structure of scientific collaboration in the journal Business & Information Systems Engineering and its predecessors

This picture shows the network of coauthorship in the scientific journals Business & Information Systems Engineering and its predecessors. As a result, you see mostly German IS researchers. The original data consists of thousands of authors, here we show only the largest connected component. Color, node layout and edge width were used to show who is working with whom.

Here is a high-resolution PDF. 

Please Email me if you have questions concerning the data or the network.

Whoooot?

There are curves of constant width besides circles and spheres

And they can be used to drill a square hole:

Little big data in German soccer

Wer gewinnt das Fußballspiel? Eine kurze Zusammenfassung aller bisherigen Bundesligaspiele. Hier ist ein hochauflösendes PDF. Tools: curl > regex > cytoscape > inkscape. Grafik von Jörn Grahl.

Econometrics vs Machine Learning

Machine learning (in particular deep learning) is currently a hot topic in CS and the social sciences because it can uncover reularities in large, high-dimensional datasets (and it really seems to work).

What is the relationship between machine learning and econometrics? I highly recommend two recent talks.

1. Hal Varian: Machine Learning and Econometrics (Youtube)
2. Sendhil Mullainathan: What Big Data Means For Social Science (edge.org)