October 28, 2019 - November 1, 2019
The development and the application of software that relies on certain mathematical and statistical algorithms is widespread in institutes of all sections of the Leibniz Association. Applications are the statistical analysis of experimental and observational data, the simulation of complex models or the optimization of engineering devices.
The second PhD Summer School of the Leibniz Network “Mathematical Modeling and Simulation” provided an introduction and an overview about two modern open source programming languages for science and statistics, namely R and Julia, which offer powerful tools for scientists. Even more, R and Julia are interoperable and one can profit from the best of both languages in everyone’s favourite language. R and Julia were presented with respect to the following criteria:
- main application areas (e.g., R for statistics, …)
- readability & learnability
- reproducibility of research
- interoperability with other open source software
- real world simulations & use as a scripting language (setup, parameter variation, …)
- sustainability in the sense of software engineering
- efficiency in time-critical applications
- parallelization of large-scale problems
The morning sessions of the summer school were devoted to tutorials on various topics. The afternoon sessions were intended to be problem-oriented and practical.
Introductory morning sessions covered the following topics:
Complex 0: Reproducibility, version control, dynamic documents
Complex 1: The R environment for statistical computing and graphics
Complex 2: The Julia programming language – design principles and challenges
Complex 3: R and Julia in action: how they help to solve mathematical/statistical problems from the applied sciences
The summer school again was problem-oriented, a concept which greatly worked also in the previous summer school 2018. The afternoon discussions were centered around problems and mathematical software pieces on which participants are actually working at. Statistical included realistic example data.
A particular thank goes to the speakers