XI. Prague R meetup

Filip Smolík (Institute of Psychology AS CR): Creating developmental norms: growth curves and their estimation using GAM(LSS) and other tools
Measurement in psychology, education, biology and other areas typically compares individuals to the distribution of the measured characteristic in the population, using e. g. percentiles. But when evaluating growth and change, for example in childhood, the value of a particular percentile will change with age. Whether it is physical height or performance on IQ tests, developmental norms must estimate curves of growth for quantiles. This is a tricky area, where inadequate shortcuts and simplifications have often been used. I will show the tools in R that can be used to find the estimates, especially tools based on generalized additive models, and point out the situations where this can be useful, as well as the conceptual problems that can arise in the process.

Jindra Lacko (VŠE): Tips & techniques for a more reproducible R environment
It is a fact of life that you are never guaranteed to get the same output from running exactly the same piece of code on two R installations at two different points in time. Both R itself and the package ecosystem surrounding it are subject to active development - bugs get fixed, old function calls are descoped to make room for new ones, and in extreme cases entire packages are delisted from CRAN. This leads to severe headaches in both academic and industry settings.
The talk will present several strategies to make your code more resilient - freezing R installation in time, using {renv} to create a snapshot of R environment, and docker images for fixing entire OS including external libraries.

Denis Abdullin (Principal Data Scientist, Johnson & Johnson): Narrator - R package for NLG with OpenAI GPT 3.5 integration
Natural language generation has been on the rise lately. Creating text narratives alongside insights in BI tools and data science reports is becoming a competitive advantage of a data scientist.
During this talk you will get to know narrator, a new open-source package for NLG, available in both R and Python. The package allows to quickly generate textual insights out of tabular datasets, highlight the most important patterns in the data, and create presentation-ready or report-ready narratives using GPT 3.5 aka ChatGPT.
We'll cover the foundations of template-based NLG systems, showcase basic and advanced use cases for narrator, and demonstrate how to use the OpenAI API for enhancing, translating, or summarizing your narratives. You'll leave this talk with a better understanding of how to use narrator to create compelling narratives that drive business impact.