Selected Topics in Data Analytics (Wintersemester 2019/2020)

Lecturer: Prof. Dr. Jürgen Döllner (Computergrafische Systeme) , Dr. Benjamin Hagedorn (Computergrafische Systeme)

General Information

• Weekly Hours: 4
• Credits: 6
• Graded: yes
• Enrolment Deadline: 01.10.-30.10.2019
• Teaching Form: Seminar / Project
• Enrolment Type: Compulsory Elective Module
• Course Language: German

Programs, Module Groups & Modules

Description

Visual Analytics "[is] the science of analytical reasoning facilitated by interactive visual interfaces" (Thomas, Cook: Illuminating the Path, 2004) it covers concepts, techniques, and processes from different scientific areas: information visualization, computer graphics, data processing, management, and statistical analysis. Here one key element of Visual Analytics is to provide users massive datasets (e.g., sensor data, financial transactions, or data about software systems structure and behavior) visually and in an interactive manner. By exploiting the strengths of the human visual system (e.g., preattentive processing and pattern recognition) structures, correlations, and patterns in massive datasets can be recognized and assessed.

Students are expected to cope with a specific aspect or technique in the area of Data Analytics, including a prototypical implementation for massive spatio-temporal datasets originating from different application areas, e.g., industrial IoT, building automatation, or health surveys.

Examples for thematic areas include:

• Integration, processing, and analysis of massive sensor datasets
• Outlier and pattern recognition in spatio-temporal data
• Hierarchy building for spatio-temporal data
• ML techniques for analytics and visualization 
• Event recognition, processing and notification
• Uncertainty and validity analysis
• …

The topics provided contain a strong relation to current research activities at the computer graphics systems group. Course material including topic presentations are available on CGS Moodle.

After successful completion of the course, many of the topics provide the possibility to extend the work into an academic publication, a master thesis, or to continue working on the topic as a working student.

Requirements

This course targets on master students enrolled in the Data Engineering program. Depending on the selected topic, machine learning and neural networks skills as well as programming skills in python or other scripting languages are of advantage.

Literature

Besides the online accessable information concerning a specific topics we will provide a set of up-to-date articles and other literature. An in depth research of further related work in the field is expected to be conducted by participants.

Learning

In general participants are expected to read up on the assigned topic and its related work. In course of the semester participants are indiviually mentored by members of the Computer Graphics Systems group. Regular meetings are scheduled individually for progress presentation and 

Examination

To sucessfully complete the cours it ist expected to 

• give a conceptual presentation that introduces the topic area, propblem statement, and presents related work
• successfully plan and implement a software development project related to the topic (50%),
• assemble foundations and results of the seminar work in a scientifically written document  (4 pages) (25%),
• give a final presentation about the topic highlightig results and specfic aspects of interest  (25%) halten.

Dates

Seminar topics are presented during the second week of the instructional period on Monday October 21st 11:00 am at H-2.58.

There is no fixed schedule during the seminar period. Meetings with mentors during the semester are scheduled individually. Appointments for presentations are announced and coordinated separately for all participants. The participants are expected to attend these presentations.

Zurück