Master Data Science

Aus Wiki
Zur Navigation springen Zur Suche springen
TODO Work in progress

The goal of ths page is to present a few possible "Tracks" through the Master Programme Data Science. A Track here is a certain collection of classes taken in order to fulfill the study plan. The idea is to document actually taken classes, not all possible choices. There could be 2-4 core areas that the study is covering mostly; this will typically cover different structures in the study plan. Ideally, the description of a track gives the list of classes taken (especially the freely elected ones, from extension modules, etc.) and short descriptions how they tie in with other lectures. Which of them are deepening understanding in what areas, which ones are comeplementing others, which of them possibly present content better than others, things like that.

Track "Machine Learning, Statistics, Data Visualization"[Bearbeiten | Quelltext bearbeiten]

The classes taken from Extension Modules and free electives were meant to cover these areas as well as possible. The overarching theme is to analyse data with statistical methods, build (machine learning) models, and visualize data in exploratory phases, as well as being able to present results using visualizations tailored to a non-technical audience.

Focus: Machine Learning[Bearbeiten | Quelltext bearbeiten]

Work in progress, links to Vowi will follow!

  • Deep Learning for Visual Computing (MLS/EX)
  • Similarity Modeling (MLS/EX)

Focus: Statistics[Bearbeiten | Quelltext bearbeiten]

Work in progress, links to Vowi will follow!

  • Multivariate Statistik, VO + UE (P. Filzmoser) (MLS/EX)

Good complement for e.g. "Advanced Methods for Regression and Classification VU" and "Statistische Simulation und computerintensive Methoden VU". A little more math, gives good introductions to methods useful for pre-processing data (dimension reduction, PCA, SVD). Also covers some robust statistics, which is a nice addition to the data scientists tool box imho.

  • Datenanalyse (P. Filzmoser) (MLS/EX) [Note: this class can only be used under certain conditions towards completion of the master.]

Introduction to more exploratory data analysis, stuff that is not covered a lot in other classes. Most students will possibly already have taken this class in their Bachelor already?

Focus: Data Visualization[Bearbeiten | Quelltext bearbeiten]

Work in progress, links to Vowi will follow!

  • Visual Data Science VU (J. Schmidt) (VAST/EX)

This class nicely complements classes "Gestaltung und Evaluation von Visualisierungen UE" and "Informationsvisualisierung VO". The latter give foundations on how to visualize data, and how to evaluate visualizations. The class Visual DS is a little more hands on, where in exercises data sets are analyzed visually and statistically, and there is the choice of creating a dashboard for a concrete dataset. This gives a deeper understanding for visualizations in more practical settings, and show how visualizations can be used. It also aims at making clear that different visualizations are useful for different audiences.

Focus: Data Fundamentals[Bearbeiten | Quelltext bearbeiten]

Work in progress, links to Vowi will follow!

  • User Research Methoden VU (G. Fitzgerald) (FDS/EX)

This class complements "Data Acquisition and Survey Methods". While the latter almost exclusively focuses on quantitative methods for analysing data from experiments, User Research almost exclusively focuses on qualitative methods. It could be beneficial to understand what kind of data such methods produce, and how they can be analyzed with data science approaches, and what tools can help with analyzing such data. Additionally, there are often mixed method studies that employ quantitative as well as qualitative methods, and understanding the latter is then necessary to fully grasp such studies.

Track "High-Performance Computing, Algorithms, Optimization"[Bearbeiten | Quelltext bearbeiten]

This is a fake place-holder entry to show that there can be multiple different tracks. Please add your own.