Freelance Data Scientist
based in Zurich, Switzerland
Freelance Data Scientist
based in Zurich, Switzerland
Alpha version of a data story mining the Arxiv metadata for insights. Teaser: Publish-or-perish attitude forces scientists to submit new papers on weekends.
Super simple app to help with Skyrim alchemy formulas. Scraped ingredient data from the Skyrim wiki and created a UI to help choosing the right ingredients. (Repo)
Beta version of the data story on how Covid-19 pandemic influenced pedestrian and bike mobility in Zurich, based on Zurich Open Data.
Data visualization done for the hands-on workshop for Open Hub Day. Collaboration with WeTechTogether. Find the slides and a recording here.
A talk about Explainable AI for ordinary humans and the symbiosis of data scientists and UX designers, given during the Swiss edition of World Usability Days 2020, themed around "Human-Centered AI".
First prize in a competition for the best data story basing on the insights from the Kaggle ML Survey 2019.
An article visualizing insights from the Himalayan Database. Who is waiting in line to summit Mount Everest? (Spoiler alert: it's the middle-aged men.)
Analysis inspired by the data collected by "The Guardian". Revenues, ownership and behavior of big oil companies reveals their incentives.
Tiny data story about the female winners of Ironman, inspired by a task for Makeover Monday.
Experimenting with different data visualizations as an answer to the challenge from "The Economist" data team.
The highlight of my PhD - we used x-rays to measure the movement of spins triggered by laser pulses. If you don't have access to the journal, here is the preprint.
3D illustration for a press release written about our research.
Read my PhD thesis to see some proto-dashboards for visual oversight of numerical modeling of scientific data. Enjoy more 3D crystal structures. Check the appendix to learn why using PCA on noisy data is a trap.
Past but not forgotten - 3D visualizations of crystal structures in organic crystals with interesting layered ordering, which I did for my Master Thesis.
What is AI and Machine Learning? Will it be a good match for your need? How to embed it in your product to maximise benefit and usability? What are the critical factors to keep in mind?
Where is the biggest value in your data? What are the analytic insights leading to immediate benefit? How to present the insights in a visually compelling and comprehensible way?
What kind of data will you need for your AI product? Which KPIs will be the most relevant for your business question? What is the baseline of current state-of-art solutions applied to your data? Which tools should you settle for?
I have been freelancing since 2019 and currently work as a freelancer with Litix. I have worked for startups and small enterprises as a consultant, and for big corporations as a contractor. My main expertise is a front-mile of data science: discovering user needs and translating them into analytical solutions, mapping the needs onto requirements for technical tools, enabling data-driven culture by creating dashboards and other analytical services. I enjoy sharing my knowledge with others and am a big proponent of collaborative innovation, incorporating elements of design sprint methodology and other interactive methods into my workflow. I have also experience in writing grant proposals. Aditionally, I am involved in teaching and supervising activities in AI-related programs at HSLU and HWZ.
Before I became an independent data scientist, I spent 2 years as a data scientist at La Mobiliere, a leading Swiss insurance company. There I worked on projects involving computer vision, NLP and data mining for improvement of internal processes. I co-authored the technical outline and timeline for the deep learning PoC, done within an international collaboration, and managed the work during the PoC as a technical lead. I communicated the results to the C-level stakeholders, which led to further joint projects.
I have a PhD in Physics from ETH Zurich. I consider having been trained as an experimental physicist as a fundament for the way I think and work as a data scientist. The most important skills I learnt are: following analytic rigor, being intellectually independent and communicating across the differences. As a physicist, I designed and performed experiments, and built theoretical models to explain experimental data. I applied statistics to real-world, biased and noisy data and developed new methods for data processing. Additionally, I participated in and managed work in international collaborations.
My basic data science toolbox includes Python (Pandas, Scikit-Learn, Seaborn, Jupyter) and Keras/Tensorflow for deep learning. For data visualization, I reach to open source solutions such as matplotlib, vega or D3.js. Among cloud providers, I am the most famililar with Google Cloud ecosystem. As many transitioners from academia, in the past I used to work with MATLAB. I also create webpages using HTML/CSS and JAM stack.
I am an avid story-teller and an enthusiast of data visualization. In the past I used to write and edit articles about astronomy, popular science and literature. Nowadays I occassionally publish on Medium and work on data stories and visual essays. I am also a member of Impact Hub Zurich and member of the Zurich organizing team of Women in Machine Learning in Data Science (WiMLDS).
Last but not least, I play music. You can hear me doing jazzy Irish folk with Ean or playing weird old instruments in trad/balfolk projects.