Statistics made easy: On using statistics to draw conclusions

You have acquired experimental data, and now you need some kind of statistical evaluation for your publication. Would like to do this without being proficient in statistical theory, and without spending too much time on it? Then this course is for you!

Date: January 16, 2025, 09:00am –12:00pm

Location: HCI G 3

Speaker: Hanspeter Schmid, FHNW School of Engineering and Environment, Institute for Sensors and Electronics

Registration: https://t1p.de/a6tk4 (for the delivery of the Python examples)

Program: This three-hour (actually 3 times 45 min. with two 15-min. breaks) course has three parts.

In part 1 (10 min.), you will be shown two easy (but correct) statistical methods that you can use for your research and publications: They base on order statistics and bootstrapping. After this part you already know WHAT you can do.

In part 2 (95 min.), you will receive a lot of background:

• Where do some of the established methods come from, and why they can be dangerous and misleading.
• Basics: Probability distributions, mean, median, percentiles.
• Credibility intervals for measurement data using order statistics.
• Robust linear regression with the Theil-Sen algorithm.
• Information theory, the normal distribution, and why we should use it sparingly.
• The difference between confidence and credibility intervals, the Bayes theorem, and why p-values are misleading.
• How bootstrapping applies the Bayes theorem properly without making you do the maths (but forcing you to write a little bit of Python code).

After this part you know WHY you can do WHAT you can do; it will give you some peace of mind every time you start to think "statistics cannot be that simple".

In part 3 (30 min.), you will work through the Python examples (Jupyter notebooks) of the course. After this block you will also know HOW you can do it, and you can easily adapt the examples to your data.

Prof. Hanspeter Schmid completed his studies in electrical engineering at ETH Zurich in 1994. This was followed by a postgraduate degree in information technologies in 1999 and a doctorate in 2000. In 2005, he began working as a research fellow at the University of Applied Sciences Northwestern Switzerland and became a full professor of microelectronics in 2012. He is also a part-time lecturer at ETH Zurich, where he gives a lecture on analogue signal processing and filtering.