Data Analysis Block Course

Name: Data Analysis Block Course
Start: 2025-07-21T09:00:00+02:00
End: 2025-08-11T13:00:00+02:00
Location: TUD/RUB/Hybrid

21 Jul 2025, 09:00 → 11 Aug 2025, 13:00 Europe/Zurich

TUD/RUB/Hybrid

TUD: Otto-Hahn-Straße 4a, 44227 Dortmund, CP-03-123 RUB: Universitätstr. 150, 44780 Bochum, NB2/158

Description

TU Dortmund: Otto-Hahn-Straße 4a, 44227 Dortmund, CP-03-123 (Campus map)
RU Bochum: Universitätstr. 150, 44780 Bochum, NB2/158
Hybrid: Zoom Link

Literature:

G. Cowan , Statistical Data Analysis
O. Behnke et al., Data Analysis in High Energy Physics

Previous courses:

RUB course, (repository)
TUD 2024

Monday 21 July
- Mon 21 Jul
- Tue 22 Jul
- Wed 23 Jul
- Thu 24 Jul
- Mon 28 Jul
- Tue 29 Jul
- Wed 30 Jul
- Thu 31 Jul
- Mon 11 Aug
- 1
  Lecture - Introduction
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
  
  lecture1.pdf
  Organisatorical matters:
  
  Exam: Please vote here for a time slot (only if you want to participate)
  
  Barbecue next Thursday: Add yourself here as a participant (and your dietary preferences)
- 2
  
  Exercises
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
  
  0_distributions.ipynb
  
  solution_0_distributions.ipynb
  
  The exercises and solutions will be provided in python as Jupyter notebooks.
  
  To have all needed dependencies you can get a python installation from Anaconda and
  
  create a new virtual environment for this course with the command
  
  conda create -n SMD2 -c conda-forge python=3.13 numpy matplotlib scipy pandas scikit-learn iminuit jacobi resample emcee jupyterlab notebook ipykernel numba-stats
  
  or instead with Miniforge where you need to replace conda with mamba in the previous command.
  
  Optional but recommended: Working with Jupyter notebooks becomes easier with the VSCode editor. Download the editor and install the "Python" and "Jupyter" extensions. You can now directly open, edit, and run Jupyter notebooks in the editor.
- 3
  
  Lecture - Error propagation
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
  
  lecture2.pdf
- 4
  
  Exercises
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
  
  1a_error_propagation.ipynb
  
  solution_1a_error_propagation.ipynb
Tuesday 22 July
- Mon 21 Jul
- Tue 22 Jul
- Wed 23 Jul
- Thu 24 Jul
- Mon 28 Jul
- Tue 29 Jul
- Wed 30 Jul
- Thu 31 Jul
- Mon 11 Aug
- 5
  
  Lecture - Information, inference and confidence level intervals
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
  
  lecture3.pdf
- 6
  
  Exercises
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
  
  1b_error_propagation_bias.ipynb
  
  solution_1b_error_propagation_bias.ipynb
- 7
  
  Lecture - Least Squares Fitting
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
  
  lecture4.pdf
- 8
  
  Exercises
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
  
  2a_least_squares.ipynb
  
  solution_2a_least_squares.ipynb
Wednesday 23 July
- Mon 21 Jul
- Tue 22 Jul
- Wed 23 Jul
- Thu 24 Jul
- Mon 28 Jul
- Tue 29 Jul
- Wed 30 Jul
- Thu 31 Jul
- Mon 11 Aug
- 9
  
  Lecture - Maximum Likelihood, Sandwich estimator, Bootstrap
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
  
  lecture5.pdf
- 10
  
  Exercises
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
  
  2b_maximum_likelihood.ipynb
  
  solution_2b_maximum_likelihood.ipynb
- 11
  
  Lecture - Markov Chain Monte Carlo, COWs
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
  
  lecture6.pdf
- 12
  
  Exercises
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
  
  3_bootstrap_mcmc.ipynb
  
  solution_3_bootstrap_mcmc.ipynb
Thursday 24 July
- Mon 21 Jul
- Tue 22 Jul
- Wed 23 Jul
- Thu 24 Jul
- Mon 28 Jul
- Tue 29 Jul
- Wed 30 Jul
- Thu 31 Jul
- Mon 11 Aug
- 13
  
  Lecture - sPlot, sFit, Unfolding
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
  
  lecture7.pdf
- 14
  
  Exercises
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
  
  4_sweights_unfolding.ipynb
  
  solution_4a_sweights_unfolding.ipynb
- 15
  
  Lecture - Unfolding, Harmonic analysis
  
  Speaker: Michael Schmelling (Max Planck Society (DE))
  
  lecture8.pdf
- 16
  
  Exercises
  
  Speaker: Jan Peter Wagner (Technische Universitaet Dortmund (DE))
  
  4_sweights_unfolding.ipynb
Monday 28 July
- Mon 21 Jul
- Tue 22 Jul
- Wed 23 Jul
- Thu 24 Jul
- Mon 28 Jul
- Tue 29 Jul
- Wed 30 Jul
- Thu 31 Jul
- Mon 11 Aug
- 17
  
  Automatic Differentiation
  
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
  
  Lecture-1A.pdf
- 18
  
  Exercises - Automatic Differentiation
  
  Exercise notebook
  
  Solution notebook
  
  week2_ex1_automatic_differentiation.pdf
  
  Install julia https://julialang.org/install/
- 19
  Machine Learning (loss, ROC)
  - Slides
  - lecture-1B-entropy.jl: introduction to entropy
  - lecture-1B-mse.jl: MSE as the simplest loss function
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
  
  lecture-1B-entropy.jl
  
  lecture-1B-mse.jl
  
  Lecture-1B.pdf
- 20
  
  Exercises - Samples, FoM, ROC
  
  Exercise notebook
  
  Solution notebook
  
  week2_ex2_samples_fom_roc.pdf
  
  Datasets for exercises
  
  Signal sample: https://cernbox.cern.ch/s/bYCCe6x1U1OM3eS
  
  Background sample: https://cernbox.cern.ch/s/0BTNgnwoiaA50A5
  
  Test sample: https://cernbox.cern.ch/s/i08RNJZJ3z8YEUX
  
  Download the samples to the same directory as the notebook, or create a link.
Tuesday 29 July
- Mon 21 Jul
- Tue 22 Jul
- Wed 23 Jul
- Thu 24 Jul
- Mon 28 Jul
- Tue 29 Jul
- Wed 30 Jul
- Thu 31 Jul
- Mon 11 Aug
- 21
  Decision Trees, Adaboost
  - Slides
  - lecture-2A-split.jl: splitting index demonstration
  - lecture-2A-trees.jl: decision tree demonstration
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
  
  Lecture-2A.pdf
  
  lecture-2A-split.jl
  
  lecture-2A-trees.jl
- 22
  
  Exercises - Decision Trees
  
  Exercise notebook
  
  Solution notebook
  
  week2_ex3_decision_trees.pdf
- 23
  Decision Trees, GradientBoost
  - Slides
  - GBDT playground
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
  
  Lecture-2B.pdf
- 24
  
  Exercises - AdaBoost, GradientBoost
  
  Exercise notebook
  
  Solution notebook
  
  week2_ex4_boosting.pdf
Wednesday 30 July
- Mon 21 Jul
- Tue 22 Jul
- Wed 23 Jul
- Thu 24 Jul
- Mon 28 Jul
- Tue 29 Jul
- Wed 30 Jul
- Thu 31 Jul
- Mon 11 Aug
- 25
  Neural Networks : General introduction
  - Slides
  - lecture-3A-xor.jl: learning xor demonstrator
  - gradient descent app: github.com/lilipads/gradient_descent_viz
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
  
  Lecture-3.pdf
- 26
  
  Exercises - Neural Networks
  
  Exercise notebook
  
  Solution notebook
  
  week2_ex5_neural_networks.pdf
- 27
  Neural networks : Bias-Variance TradeOff and regularization techniques
  - Slides: Lecture 3 pdf contains but A,B parts
  - lecture-3B-UAT.jl: Universal approximation theory demonstration
  - lecture-3B-NMIST.jl: training of MNIST classification
  - playground.tensorflow.org
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
  
  lecture-3B-NMIST.jl
  
  lecture-3B-NMIST.pdf
  
  lecture-3B-UAT.jl
  
  Lecture-3.pdf
- 28
  
  Exercises - NNs on example data
  
  Exercise notebook
  
  Solution notebook
  
  week2_ex6_neural_networks.pdf
Thursday 31 July
- Mon 21 Jul
- Tue 22 Jul
- Wed 23 Jul
- Thu 24 Jul
- Mon 28 Jul
- Tue 29 Jul
- Wed 30 Jul
- Thu 31 Jul
- Mon 11 Aug
- 29
  Generative networks, Normalizing flow
  - Slides
  - lecture-4A-transformations.jl: demonstration of bijectors
  - lecture-4A-RQS.jl: tuning RQS
  Speaker: Prof. Mikhail Mikhasenko (Ruhr Univeristy Bochum)
  
  lecture-4A-RQS.jl
  
  lecture-4A-transformations.jl
  
  Lecture-4.pdf
- 30
  
  Exercises - Normalizing flow
  
  Notebook NF basics
  
  Notebook NF training
  
  week2_ex7_normalizing_flow_basics.pdf
  
  week2_ex8_normalizing_flow.pdf
- 31
  
  BBQ
  
  Sign up here: https://hackmd.e5.physik.tu-dortmund.de/3XzYG8C4ScuKaNh43QB3GQ?both
  
  Location: https://maps.app.goo.gl/qwZvu8eUB14JzKTw5
Monday 11 August
- Mon 21 Jul
- Tue 22 Jul
- Wed 23 Jul
- Thu 24 Jul
- Mon 28 Jul
- Tue 29 Jul
- Wed 30 Jul
- Thu 31 Jul
- Mon 11 Aug
- 32
  Written Exam (in person, in Dortmund and Bochum)
  - Pen and paper, calculator (non-programmable)
  - Time: 1:30h
  - 5-6 Problems
  - ChatGPT is not allowed
  ML_part.pdf
  
  SMD2_part.pdf

Choose timezone

Data Analysis Block Course

TUD/RUB/Hybrid

Datasets for exercises