Lab / PG - LLMs for Statistical Data Analysis

Statistical data analysis is often hard, time-consuming and prone to errors. Recent advances in LLMs raise the question of whether they can support this process and achieve comparable or even better results compared to human analysis.

The task of this PG/Lab is to evaluate the capabilities of LLMs for statistical data analysis. You will review published papers and use the corresponding anonymized datasets to reproduce the reported analyses, using LLMs. The generated results will then be compared with the original findings reported in the publications. In cases of discrepancies, you will manually verify the results.

Since this project focuses on statistical data analysis, a basic understanding of statistics is required. Suitable for group work.