Contract type: consulting
Client: Upper Austrian Insurance
Corporation
Role: Data Science Consultant
Time period: Jul 2015 –
Nov 2015
Volume: 125hrs
data science • Linux • Python • numpy • scipy • matplotlib • ipython • scikit-learn • Kyoto Cabinet • LevelDB
Based on internal datasources from within the Upper Austrian Insurance Corporation, I extracted descriptive statistics and conducted automated pattern recognition through machine learning methods as well as predictive modelling. I presented the results directly to C-level executives.
I was given raw data from pre-existing IT infrastructure. In order to bring the data into a homogeneous format suitable for analysis, I developed a great deal of preprocessing code.
I analysed the data through a Bayesian approach to multivariate analysis, i.e. I analysed the data in terms of combinations of variables, rather than looking at variables in isolation.
As part of the project, I was able to extract a considerable amount of actionable insight, as well as fitting a predictive model for immediate use by the client.