Loading
00EINEU010 20W 5SWS PR Python for Engineering Data Analysis - from Machine Learning to Visualization   Hilfe Logo

LV - Detailansicht

Wichtigste Meldungen anzeigenMeldungsfenster schließen
Allgemeine Angaben
Python for Engineering Data Analysis - from Machine Learning to Visualization 
00EINEU010
practical training
5
Winter semester 2020/21
Associate Professorship of Simulation of Nanosystems for Energy Conversion (Prof. Gagliardi)
(Contact information)
Details
Allocations: 1 
Angaben zur Abhaltung
THE COURSE IS ALSO TAKING PLACE IN THE WINTER TERM 2020/21! Sign up should be possible latest from 2020-10-16.

With the success of data-centric business models often built around machine learning and AI, data science and modeling have been pushed into the focal point of various disciplines, reaching from factory automatization through business analytics and materials science.
Today, the statistical data analysis concepts associated with the resulting “data science paradigm” are increasingly introduced as a “new way to do things” to all scientific and engineering domains, which have historically already relied on simple and predictable analysis methods (e.g. linear regression and 2D-plots). Unfortunately, often the sizable heritage of the “new” methods in engineering (see Forrester et. al., 2008) is overlooked and new methods are applied uncritically.
In this regard, this lab course aims to introduce the Python programming language as a tool for engineers and scientists to perform all the data analyis tasks typically arising in a lab (data cleaning, calculating statistics, linear modeling, visualization); in addition the relation of simple statistical modeling to machine learning methods and their inherent shortcomings will be presented.
If a students wants to learn to program efficiently in python, EI0508 “Projektpraktikum Python” is more suitable; for the ones interested in a focused introduction of the statistics behind machine learning, they should take EI04016 or a course from the mathematical department.
A rough outline of the lab course could be:
(0) Kickoff + Motivation
(1) Introduction to the Python programming language
(2) Accessing data – interacting with lab equipment, databases, MS Excel and the internet
(3) Data Visualization – do’s and don’t’s
(4) Modeling data – simple statistical techniques to create a data model
(5) (Statistical) Learning – can a data model be used for prediction?
None, a previous course on programming or algorithms is recommended (and included in the core EE-curriculum with module IN8009)
After successful participation in this course, students
- know how to solve algorithmic problems in Python
- are able to independently perform basic data analysis/statistics tasks on datasets
- have a solid understanding of the available tools for 2D visualization (including images) and be able to start working with 3D-data and associated visualization tools (might be further explored in the project)
- have a conceptual understanding of various methods (linear regression, regularized regression, neural networks) to model (high-dimensional) datasets and the relation to basic statistics
- know about the limitations of said models in prediction and optimization settings
- have experience with state of the art tools for building data-analysis workflows using Python (jupyter, pandas)
English
laboratory
The students will solve basic data analysis problems during regular hands-on tutorial sessions in the computer room (twice a week). This part will be supported by slides and ressources on Moodle. To ensure steady progress, additional exercises are given for them to solve by self-studying.

An internet-accessible server will be provided, where students can use Jupyterhub or a classic editor to do their programming assignments.
Details
Für die Anmeldung zur Teilnahme müssen Sie sich in TUMonline als Studierende*r identifizieren.
Zusatzinformationen
- Python 3 – Reference documentation [https://docs.python.org/3/]
- Sweigart, Al: Automate the boring stuff with Python [https://automatetheboringstuff.com/]
- Forrester et. al.: Engineering Design via Surrogate Modelling: A Practical Guide, Wiley (2008)

additional, specialized literature on specific problem sets will be provided during the course.
Online information
e-learning course (moodle)