Event
Data Science for the Industrial Internet of Things [SS222540493]
Lecturers
Organisation
- Karlsruhe Service Research Institute
Part of
- Brick Seminar in Business Administration A (Master) | Industrial Engineering and Management (M.Sc.)
- Brick Seminar in Business Administration B (Master) | Industrial Engineering and Management (M.Sc.)
- Brick Seminar in Business Administration A (Master) | Economics Engineering (M.Sc.)
- Brick Seminar in Business Administration B (Master) | Economics Engineering (M.Sc.)
- Brick Seminar in Business Administration A (Master) | Information Systems (M.Sc.)
- Brick Seminar in Information Systems (Master) | Information Systems (M.Sc.)
- Brick Seminar in Business Administration A (Master) | Information Engineering and Management (M.Sc.)
- Brick Seminar in Business Administration A (Master) | Economathematics (M.Sc.)
- Brick Seminar in Business Administration B (Master) | Economathematics (M.Sc.)
Note
Learning Objectives
- Gain practical experience in translating a business problem into a data modeling problem
- Apply solid theoretical foundations from lectures to real-world data
- Acquire hands-on experience with industrial data science tools
- Learn how to communicate data science findings to business stakeholders
Course Credits
The practical seminar can be credited as Seminar Betriebswirtschaftslehre A [WIWI-103474] (3 ECTS). Other courses can be credited upon request.
Seminar Description
The Internet of Things is significantly transforming industries such as automotive, healthcare, and energy. With the rise of ubiquitous computing power, internet access, and economical sensors – physical products turn into cyber-physical smart products that create vast amounts of data.
Current airplanes for example have around 6.000 sensors, creating around 1 TB of data per flight. This data is about the size of all tweets in 3 months worldwide. And this number is growing tremendously. But only 3% of potentially useful data is tagged today, end even less is analyzed. Although Internet of Things use cases such as predictive maintenance are projected to help companies save $630 billion by 2025 (McKinsey, 2015), companies struggle to turn sensor data into actionable insights. To solve this challenge, substantive expertise needs to be combined with skills from software engineering and statistics and machine learning to generate valuable insights from machine data.
The practical seminar is held in cooperation with industry partners of the KSRI, which provide some real-word datasets. Students will then work in teams of three in a close and agile collaboration with the industry subject matter experts from around the world, making use of to the CRISP DM methodology (Chapman et al. 2000)
There will be four different topics and datasets, each assigned to a team of three students. The assignment will be done in the kickoff in calendar week 18. The exact date of the kickoff event will be determined when the participating students have been selected. Attendance at the kickoff event in calendar week 18 is mandatory and a prerequisite for participation.
Expertise in Python and Data Science / Machine Learning is strongly recommended.
Contact
Dominik Martin – dominik.martin@kit.edu
Dr. Niklas Kühl – niklas.kuehl@kit.edu
The practical seminar will be held in English. Application documents can be handed in in English or German.