by Greg Kotkowski

A week ago I started my 3-months secondment at CERN. This is the first time that I have ever come to CERN. Don’t blame me, I’m a statistician, not a physicist, so I have never had neither an opportunity nor any need to come here.

So far I’m totally overwhelmed and knocked out by CERN. Seeing all those researchers and other smart people in one place is an extraordinary experience. Actually, I feel kind of stupid in their presence, so I should always bear in mind that I’m still an EARLY stage researcher.

The whole scenery around CERN is also astonishing. The swiss Alps look fabulous. I’m also very surprised by the people who live here. Every stranger that I meet on the streets says “Bonjour!”. It is so simple, but makes me like and trust the local community here. Even the French police is friendly: when they stopped me for driving with one of the frontal lights burned out they just told me to replace it at the next occasion.

As CERN is so big and lots of researchers are coming in and out, it is easy to get lost. I’m very thankful to Tommaso Dorigo, who came here for two days to introduce me to researchers and other important people with whom I could possibly work and benefit from. Thanks to Tommaso I now have tools and a good point to start from.

The software used at CERN is also a challenge for me as a statistician. Data stored in root files is not straightforward to be analysed for me, not to mention the parametrisation of selectors or some other routines run in root. However, thanks to the help and the experience of Pablo I’m finally able to work on the data using Python or even R (that I’m more familiar with).

My main goal for the time here is to check the statistical properties of the Hemisphere algorithm recently developed by Tommaso and others. The algorithm aims at estimating the background density by smearing the signal fraction in the input data.

On the other hand, the algorithm has to be sensitive enough and not remove the important features of the background. The flexibility of the model is handled by some parameters, but their influence on the output is not yet known. I’m very excited to study the algorithm’s properties and I hope there will be a chance for significant optimisation.

Soon I’ll give you on overview on my work progress. For now I’m going to discover the rich scientific life of CERN.