Pandas Summary

In any data-analysis workflow, there's one move you make before any other: what am I actually looking at? Pandas Summary is that move. It runs describe() on a DataFrame and returns the whole statistical overview - count, mean, standard deviation, min, the 25/50/75 percentiles, and max - for every numeric column, in a single table.

It's the pack's "big picture" node, and it's genuinely the one to reach for first. Before you sort, filter, or plot anything, wire your freshly loaded frame into Pandas Summary, display the result, and you instantly know: which columns are numeric, which are full of holes (look at the counts), how spread out everything is, and whether anything's wildly out of range. The author's baseball tutorial walks exactly this path - load the data, summarize, then decide what's worth extracting. Every real workflow in this pack should have one of these near the front.

How it works

It's a one-liner over pandas: dataframe.describe(). That's the same function you'd call in a notebook, so the output is the standard summary table pandas generates - descriptive stats for numeric columns, with non-numeric columns skipped entirely. There's no configuration here; you feed it a frame and get the whole report. Note the result is a DataFrame (rows = stats, columns = your original columns), so unlike the Pandas Std/Pandas Var/Pandas Sum nodes that return a Series, this one stays table-shaped.

Inputs and outputs

• dataframe (DATAFRAME) - the frame to analyze.

Output: a single DATAFRAME - the summary table. Wire it into Pandas Show DataFrame or Pandas To String + Pandas Show Text to actually read it, or Pandas To Numpy if you want the stats as an array for further math.

Installing this pack

This node ships in ComfyUI-Data-Analysis by Hide Inada (HowToSD). CPU-only, no GPU, no model downloads - but the pack needs its Python stack, which stock ComfyUI lacks: pandas, matplotlib, seaborn, scipy, scikit-learn, openpyxl, lxml.

ComfyUI Manager: Manager → Custom Node Manager → search "Data analysis" → install ComfyUI-Data-Analysis → restart ComfyUI and refresh the browser. Manager installs the deps.

Git clone:

cd ComfyUI/custom_nodes
git clone https://github.com/HowToSD/ComfyUI-Data-Analysis.git
mv ComfyUI-Data-Analysis data-analysis   # example workflows depend on this folder name
pip install -r data-analysis/requirements.txt

Restart ComfyUI after installing.

Troubleshooting

• Only some columns show up. describe() only summarizes numeric columns; text columns are skipped. If you expected a column and it's missing, it's non-numeric - convert it or accept that categoricals don't get a mean.
• Counts look lower than the row count. That's describe() telling you about missing values. Each column's count is how many non-NaN values it has; big gaps there are worth investigating before you analyze.
• The table is wide and clipped. Resize the Pandas Show DataFrame node, or route through Pandas To String and read the text in a bigger Pandas Show Text widget.