PandasPlotBench is a 175-task benchmark showing that LLMs generate good Matplotlib and Seaborn plots from brief instructions but struggle with Plotly.
Le veraging large language models for data analysis automation,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.SE 1years
2024 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Drawing Pandas: A Benchmark for LLMs in Generating Plotting Code
PandasPlotBench is a 175-task benchmark showing that LLMs generate good Matplotlib and Seaborn plots from brief instructions but struggle with Plotly.