My name is Tim Hagmann, and I am the designated Director at LUSTAT, an independent public organization based in Luzern, Switzerland. LUSTAT ist dedicated to delivering comprehensive data and precise analyses for the canton and city of Lucerne, as well as six other cantons of central Switzerland. My team and I will develop intelligent data products that leverage both internal and external data sources. Our goal is to provide actionable insights for authorities, administrations, research institutions, and businesses through the application of various data science methods.
I'm also politically active and an elected member of the state council (Landrat/Kantonsrat) in the canton Basel Country for the Green Liberal Party. In that capacity I'm a member of the Economic and Health Commission (VGK) and a substitute member of the Environmental Protection and Energy Commission (UEK) as well as on the advisory board for the organization Regio Basiliensis. See my political website for more information.
I love reading and traveling. During my studies I spend a considerable amount of my study time abroad through two international exchanges (one year at the University of Limerick, Ireland and 6 months at the Université de Lille, France). I also did NGO work in the SOS Children Villages in Ethiopia and I speak German, French, English and a little bit of Spanish.
I believe in the importance of a livelong education (learning as well as teaching). I'm holding an economics degree from the Universität Zürich and a statistics degree from the Université de Neuchâtel. More recently I finished a part-time two-year data science program at Harvard. I also teach business and economics during economy weeks to high school students.
with links and catchy titles
Mar 2015 - Jul 2017
The Active App is designed to help Sanitas customers to get more exercise and eat healthily. It counts steps and tracks a users activity while cycling or swimming. It uses various nudging mechanisms to encourage its users to maintain a healthy lifestyle.
Jul 2015 - Mar 2015
Baloise Plus is the new benefits program for Baloise customers. The more contracts customers are holding the larger their benefits. I was the lead data scientist on the project.
Aug 2014 - Jan 2015
The interactive risk map for burglary and theft is a new tool for Baloise customers, insurance agents and actuaries to analyse the individual risk distribution at a cantonal as well as municipal level. It is build based on internal data, external data from the federal office of statistics and on top of the D3.js framework.
Feb 2012 - Jun 2012
The paper examines the classification and regression tree (CART) algorithm as a tool to facilitate the process of model specification needed by GLM's when modelling premiums, i.e., to use CART for variable selection, interaction detection and preliminary understanding of the data.
Aug 2009 - Mar 2010
In the paper published in "Labour" we examined how different student employment statuses during tertiary education affect short-term and medium-term labor market returns. Using a representative survey of Swiss graduates of tertiary education, we found significant positive labor market returns of ’earning while learning‘, but only for related student employment and not for unrelated student employment.
Apr 2009 - Sep 2009
The paper examines the short and long run consequences of the Irish Famine of 1845 to 1850. By adopting Krugman’s (1991) two-region model to the Anglo-Irish case, it argued, in line with O'Rourke and other cliometric economists, that the Famine constituted such a shock on the economic fabrique of Ireland that the following Irish industrial backwardness was a result of it.
Professional Graduate Data Science Coursework
Jan 2016 - Jun 2016
The course topics included good design practices for data visualizations, methods for visualizing from different sources, and programming of interactive web-based visualizations using D3 (Javascript). All the code to reproduce the results below is on Github.
Aug 2016 - Dec 2016
The course was the first part of a series about advanced data science methods. Topics include the analysis of high dimensonal data with linear methods such as lasso and ridge regression, bayesian modeling and sentiment analysis as well ensemble methods such as bagging and boosting. Most of the programming was done in Python. All the code to reproduce the results below is on Github.
Jan 2017 - Jun 2017
The course is the second part of a series about advanced data science methods. Topics include the use of non-linear statistical models such as smoothers and GAM, unsupervised learning, and deep learning with neural networks. The major programming languages are R and Python. All the code to reproduce the results below is on Github.
Aug 2017 - Dec 2017
The emphasis of this course is on mastering different important big data technologies. The focus lies on Spark, Tensorflow and Streaming technologies which allow analysis of data in flight, that is, in near real time. Furthermore the so-called NoSQL storage solutions exemplified by Cassandra are examined. An additional focus lies on memory-resident databases and graph databases (GraphX and Neo4J) and scalable messaging systems like Kafka and Amazon Kinesis. All the code to reproduce the results below is on Github.
Hit me up at any of the links below.