Lectures on Data Science:

Time Series Analysis

Thursday, 16.09. - Friday, 17.09.2021 · 9 a.m. - 5 p.m.

The course aims to teach you the basic knowledge of time series analysis with several hands-on sessions. On the first day, you will learn about stochastic models and on the second day, we will deal with deterministic effects (trends and periodic patterns).


Day 1: Stochastic Models

  1. Autoregressive models
  2. Markov Chains
    1. Short Introduction
    2. Application to wind power time series
    3. Discussion: The problem of correlations 
  3. Kramers Moyal expansion
    1. Not-so-short introduction
    2. Application to power grid frequency time series: estimation of essential power system parameters such as the effective primary control.
    3. If time allows: Application to a Geophysical Time Series.

Day 2: Deterministic Patterns

  1. Trends
    1. Intro: When is a trend significant?
    2. Application to annual wind speed time series
    3. Discussion
  2. Periodic patterns
    1. Intro: The Fourier transformation
    2. Application to a sample time series.
    3. Discussion Limitations of Fourier analysis for short and noisy time series. 
  3. Alternatives to Fourier analysis for short time series such as SSA or multi-taper methods

Requirements:  For the hands-on session, please bring your laptop.