Don't Jam the LHC: A causal analysis of Code Jam data

Speaker:
Carlo Alberto Furia, Università della Svizzera italiana

Abstract:
There is abundant observational data in the software engineering domain, whereas running large-scale controlled experiments is often prohibitively challenging. Fortunately, modern causal analysis techniques provide means to construct statistical models of observational data that are more likely to capture genuine causal relations – as opposed to mere correlations. In this talk, I will demonstrate some of these causal analysis techniques on data from the Google Code Jam programming competition. We will see that the estimates of the programming languages’ impact on the results in this competition change considerably depending on whether we take causal relations into account.

Biography:
Carlo A. Furia is an associate professor in USI’s Informatics Faculty, and a member of the Software Institute.

Chair:
Marco Raglianti