Master s thesis topics in Data Science Gaming Parallel Programming Aske Plaat, Thomas Bäck, Joost Kok, Jaap van den Herik, Arno Knobbe, Siegfried Nijssen, Walter Kosters, Fons Verbeek, Michael Lew, Michael Emmerich, Jeroen Laros, Matthijs van Leeuwen, Jan Adriaanse, Jeroen van der Leijé, Anske van der Bom
Near-future Prediction of Demand for Flowers What: A big data modeling approach to the european demand for flowers. Factors involved are historic data, weather data, flower sales, etc Profile: Data Mining Advisor: Nijssen Start: Now, discuss with company in Rijnsburg, internship
United Nations Global Pulse What: Analysing social network data (twitter, etc). Find correlations, analyse phone data, text mining for the United Nations Profile: Data Mining Advisor: Nijssen, Baar (Center for Innovation) Start: discuss with United Nations Global Pulse, internship
Decision support at Intensive Care What: Create a decision support tool for Intensive Care doctors. Use as inputs 10 years of high frequency measurements, and patient data. User interaction. Profile: Interaction modeling, Data Mining, time series data Advisor: Nijssen, Cachucho, Van der Bom (LUMC) Start: Now
Data Analytics Aegon Customer Intelligence What: Pattern finding in customer data at an insurance company. Also: fraud detection Profile: Data Mining Advisor: Kowalczyk Start: Now, internship
Video Analysis of Parent Child Interaction What: Analyse video data of parent child interactions. From simple (smiling) to complex (reaction, empathy). With Institute of Education and Child Studies (prof. Mesman). Profile: Image analysis Advisor: Verbeek, Lew Start: Now
Behavior Mining in Casino Data What: At a Casino data startup, model and analyze data and customer behavior. What factors make people gamble? In collaboration with Faculty of Social Sciences. Profile: Data Mining & Behavior Advisor: Takes, Plaat, Bücker Start: Now, internship
Sportsdata analysis What: With sportsdata companies and IOC/NSF, analyse behavior data of athletes, or predict sports outcomes, or transfer values Profile: Data Mining & Behavior Advisor: Kok, Knobbe, Cachucho Start: Now
KLM Flight Safety pilot data analytics What: At KLM Flight Safety study pilot training data, pilot operational data, and in-flight data for performance analysis (Germanwings) Profile: Data Mining & Behavior Advisor: Plaat, Boer, Van Leeuwen Start: Now, internship
Fuel efficient Shipping What: At a medium sized shipping firm, combine loading, tracking, fuel consumption, weather data to optimize logisitcs and to optimize fuel consumption. Collaboration with TNO Profile: Data Mining, Time series analysis Advisor: Cachucho, Nijssen, Knobbe Start: Now, internship
Mapping the Genome of Business What: Based on high resolution data from DNB, CBS, and Exact, map the genome of business: what makes companies do well, can you predict crisis events? In collaboration with Rotterdam School of Management (prof. Vervest) Profile: Data Mining & Time series analysis Advisor: Van Leeuwen/Nijssen, Plaat, Vervest Start: Now
Gephi for Gamespace: Visualizing Complex or High Dimensional Spaces What: Chess programmers and data miners have one thing in common: they see their search space and feel the need for an algorithmic change. By writing a flexibile multi purpose visualization tool their intuition will be improved considerably. We can take existing problems, games, or you can use your own. Profile: Visualization, Artificial Intelligence, Games Advisor: Van Leeuwen, Plaat, Nijssen, Kosters Start: Now
Learning from perfect knowledge in games What: End game databases exist that contain perfect solutions for board situation in chess, checkers, othello, and other games. Can we generalize patterns to learn usable knowledge from these databases? Relation with Minimum Description Length of Van Leeuwen/Grünwald. Profile: Artificial Intelligence, Mathematics, Gaming Advisor: Van Leeuwen, Kosters, Plaat Start: Now
GPGPU Scheduling of irregular parallelism What: GPUs are highly suited for regular data parallel programming styles. Most real world problems are irregular in nature. This topic explores scheduling of irregular task parallelism on GPGPUs. Profile: Parallel Programming, CUDA Advisor: Mirsoleimani, Stefanov, Plaat Start: Now
Solving Leiden s Traffic Issues What: The city of Leiden is full of sensors. Traffic data can be used to improve Leiden s traffic situation. Collaboration with traffic experts of the city of Leiden Profile: Data Mining, Time Series Analysis Advisor: Van der Leijé, Knobbe Start: Now, consult with Leiden city hall
The Pulse of the Organization What: A large set of employee survey data is waiting to be analyzed to predict the success of reorganizations at large companies in the energy and financial sector. In collaboration with prof. Adriaanse of Turnaround Management. Profile: Data Mining & Organization Advisor: Adriaanse, Plaat Start: Now, internship