Work

Take a look at a selection of projects I worked on either in academia, business or teaching. Most of these projects were conducted in teams. More information about the context and co-authors upon request.

Business

Project Description	Research Focus	Data type	Methods used	Status
Retention Modeling	Conceptualizing and developing a retention model and ETL pipeline for a non-contractual mobility setting.	Transactional	Simple heuristics, linear models, CDK, AWS Batch, GitHub Actions, Docker	on-going
Dynamic Pricing	Maintaining and enhancing a dynamic pricing microservice for a mobility provider.	Transactional	Test-based Data Science, Non-linear models, AWS Batch, Docker, MLFLow	on-going
Brand perception of energie provider	Survey-based research on a multitude of brand-related mesaures and customer satisfaction.	Survey	Linear models, Principal Component Analysis, Cluster Analysis	final
Success drivers of brand space event	Survey-based research on the perception of a temporary live marketing campaign and its effects on the brand perception for an insurance company	Survey	Linear models, Principal Component Analysis, Cluster Analysis	final
Predict cart abandonment	Based on sales data, we predict the probability for a customer to abandon their shopping cart via an artificial neural network.	Sales data	ANN, Keras & Tensorflow	final

Scientific

Project Description	Research Focus	Data type	Methods used	Status
Impact of 9 € Ticket on private mobility provider	Estimating the effect of the government subsidized 9 € Ticket on the customer base of a private mobility provider.	Survey + Transactional	Propensity Score Matching	final
Short and long-term effect of bundle promotions	Based on actual sales data, we investigate whether customers react towards bundle promotions with higher number of store visits, higher quantity purchased and higher spending.	Sales data	Linear & logit models, Maximum Simulated Likelihood	on-going
Value of customer relationship depending on acquisition type	Based on actual sales data, we look at the value of a customer relationship on multiple levels. We specifically investigate differences in key customer behavior (e. g. repeat purchase, disount-proneness, probability to return an order) depending on wether they were acquired by a discount, or not.	Sales data	Network analysis, linear & logit models, simulation	on-going
Experiment: Response to product bundles	We systematically varied different types of product bundle framings and measured customers’ response for our focus brand against a competing brand.	Survey, eperimental	logit & count model, Maximum Simulated Likelihood	on-going

Teaching

These projects were either conducted under my supervision or created by me to teach and illustrate methods in class.

Project Description	Research Focus	Data type	Methods used	Status
Experiment: Effect of premium promotions	Measure the response (brand choice probability & attractiveness) of subjects towards different premium promotions.	Survey	Linear & logit models, Structural Equation Modeling	final
Media consumption behavior	Survey-based research on the media consumption behavior, media spending and technology usage.	Survey	Linear & logit models, Principal Component Analysis	final
Qualitative: Money-back-guarantees	Explore what might impact customer behavior in light of fast-moving consumer good’s money back guarantees.	Qualitative	Focus group analysis	final

Fun

I sometimes do silly stuff to increase my skill set or learn new tools.

Project Description	Research Focus	Data type	Methods used	Status
Radio playlist	Web scraping and analysing my favorite radio stations playlist over one year.	Scraping	Web scraping, exploratory analysis	fun
Urban gardening	Monitor and analyze my balcony plants’ hydration level via sensors and RaspberryPi.	Sensor data	RaspberryPi & Python	fun
Getting rich	Create a Python trading robot that utilizes Poloniex API to automate micro crypto transactions.	API data	Polonies API & Python	still not rich