Curdin Derungs,

MSc. Geographic Information Systems UZH, Dr. sc. nat UZH, DAS Applied Statistics ETH

Project Manager Maschine Learning at ehub and NEST

Google Scholar, LinkedIn

After 10+ years in the Data Sciences, my fascination with exploring new data is still the same. The special moment when a pattern slowly starts to emerge from the initial randomness. In different functions and organizations I have worked for several years with spatial data, textual information and time series.

Time series analysis: First at the Lucerne University of Applied Sciences and Arts and now at Empa, I conceptualize and lead interdisciplinary industry projects for the development of automations and algorithmic optimizations in the field of buildings and districts. Measurement series form the data basis. Often the hundreds of millions of measured values originate from dozens of sensors and in their combination result in a pattern. In the course of this work I attended a postgraduate course (DAS) in statistics at ETH. I wrote the thesis on predictive deep learning of energy consumption data.

NLP: I wrote a PhD thesis on data mining in large digitized text data at the University of Zurich (UZH). The goal was to automatically identify spatial locations in unstructured text data and thus make the spatial references in texts explicit. As tools I used explorative analysis methods (e.g. text clustering like topic modeling), as well as classification methods (e.g. word embedding like Word2Vec). After the PhD I was allowed to lead a small service group at the university. The group supported researchers in the social sciences and humanities in the analysis of increasingly digital and large data. Examples of digital textual data included historical books, social media posts, or blog entries.

Spatial Statistics: After my master studies in Geographic Information Science at UZH, I worked for a reinsurance company in statistical modeling of natural hazards and then for Grün Stadt Zürich (GSZ). At GSZ I supported the 300 employees in collecting, storing and visually processing spatial information.