ComfyUI-Data-Analysis

The node that types a float so you don't have to

Hand a typed integer to any node in the pack

The glue that turns typed JSON into pack-ready input

Actually see the number instead of trusting the wire

See your integer before you build anything on top of it

Turn a text field into a proper string node

The pack's multiline text source, ready to wire

Get a real bar chart out of a DataFrame and into your workflow

Trends from your DataFrame, rendered as a real image

Proportions at a glance, straight from a DataFrame

See the relationship between two columns at a glance

Type a list, get a float32 array

Build an int32 array by typing a list

See what's actually in that array before it breaks your workflow

Strip the useless size-1 dimensions out of your array

Add two DataFrames cell by cell — alignment included, no matrix math

Shift every value in a DataFrame by one constant

Add one integer to every cell — the non-float cousin

Shift each column by its own amount, using a Series

Force every cell to a number — junk becomes NaN, nothing explodes

Make the whole frame integer — with NaN still allowed to exist

Flatten the whole table to text in one pass

Pluck one datetime out of a table by its row and column labels

Grab a single number out of a table and wire it anywhere

Pull one integer out of a DataFrame by label

Write a datetime into a specific cell — and know it mutates

Patch one cell with a number, in place

Write one integer into a cell and hand the frame back

Drop a string into a single cell, in place

The node that turns one table cell into prompt text

Filter a table down to the rows that matter, with a True/False Series

Stop guessing what your columns are called

The whole correlation matrix, in three flavors

Element-wise cosine — for when your data is actually an angle

Non-null counts per column, in one Series — the data-quality check

Covariance, not correlation — the matrix that keeps the scale

The fastest on-ramp to pandas in ComfyUI

Build a DataFrame from a Python dict, the way ComfyUI can actually wire it

When your row labels actually mean something

Up to ten columns, one dict per column

Hand your arrays to pandas without the round-trip

Turn a PyTorch tensor into a table you can analyze

One-dimensional data, dict-shaped

The simplest way to get a Series onto your canvas

A Series with labels you actually chose

When your CSV's first column is an index, not data

Count how often two categories line up, in one node

Track the running maximum down a column

The running minimum, for when only the worst-so-far matters

Running products, with a warning about compound explosions

The running total, the workhorse of the cumulative family

Divide one DataFrame by another, alignment and all

Scale a whole table by one number

Integer-scaled division when you know the number is whole

Divide every column by one Series, the normalization move

Sweep the repeat rows out of your table

Delete the rows with missing values before they corrupt your math

The == operator of your data graph

Flag every cell that equals exactly this number

The node for matching integer codes

Find out what's actually inside that .xlsx

E^x across your whole table, with strings safely out of the way

The ML-prep node that collapses four nodes into one

Ship your table straight to PyTorch tensors

Patch missing values before they poison your math

Fill the gaps with a whole number, not a guess

Element-wise 'greater than or equal to' between two frames

The threshold filter you'll actually use daily

Integer thresholds for when decimals don't matter

The 'group by year, sum the hits' node the whole pack is built around

The strict 'greater than' comparison, cell by cell

Strict 'greater than' against a threshold

Strict integer thresholds for counts and ranks

A quick peek at your first rows, ComfyUI-style

Glue two tables side by side

Cut a wide table into two narrower ones

Pull one datetime cell out of your table

Extract one number from your table to drive a node

The node that turns a table cell into a number you can actually use

Poke a datetime into any cell of a DataFrame

Write a float into one cell without touching the rest

Correct a single integer cell, position by position

Overwrite one text cell, the un-pandas way

Grab one cell out of a DataFrame as text

Select rows by a list of positions, not a range

Pull one row out as a Series, by its position

Slice a run of rows out of a DataFrame

Get a DataFrame's row labels out where you can see them

Turn a DataFrame's index into text you can actually read

Map every missing value in your data before it bites you

Merge two DataFrames on a shared column, visually

How peaky is your data? Ask kurtosis.

Element-by-element 'less than or equal to' between two tables

Flag every cell that's at or below a float threshold

Flag every cell at or below an integer cutoff

Load a CSV into a DataFrame — the node most of this pack starts from

The CSV loader for files that aren't UTF-8

Load a CSV that already has an index column

Read an Excel sheet into a DataFrame, no export step needed

Scrape the tables out of an HTML page, up to ten at a time

The on-ramp for tabular data in ComfyUI

Pull one cell out of a DataFrame as text

Picking rows when your index has two keys

One whole row, by label, as a Series

Natural logs on a DataFrame, with the mess cleaned up

Element-wise 'less than' between two DataFrames

'every value under this number?' in one node

The quick 'values below this threshold' mask

The per-column ceiling of your table

The average of every column, in one Series

The average that survives your outliers

The reshape that turns wide data into long data

The floor of every column, fast

The most common value in every column — all of them

Multiply two DataFrames, cell by cell

Scale an entire DataFrame by one number

Multiply a whole DataFrame by one integer

Multiply a DataFrame by one row's worth of values

Find where two DataFrames disagree, cell by cell

Flag every cell that isn't this number

Pandas Ne Scalar Int

Pandas Pivot

Pandas Pow

Pandas Pow Scalar Float

Pandas Pow Scalar Int

Pandas Pow Series

Pandas Rename

Pandas Rename Advanced

Pandas Replace

Pandas Replace Advanced

Pandas Save CSV

Pandas Save JSON

Pandas Select Column As Series

Pandas Select Columns

Pandas Select Rows

Pandas Series To DataFrame

Pandas Series To String

Pandas Set Index

Pandas Show DataFrame

Pandas Show Index

See what's actually inside your Pandas Series

The 'just let me see it' text viewer every workflow ends with

Apply sine to a whole DataFrame, non-numeric cells become NaN

Is your data lopsided? One node for skewness

Sort a DataFrame by a column, ascending or not

Standard deviation per column, in one node

Turn datetime columns into pretty strings

Subtract one DataFrame from another, cell by cell

Subtract one float from an entire DataFrame

Subtract a whole number from every row

Subtract a Series from every row of a DataFrame

Column totals, as a Series — the SQL SUM of the graph

The first node you should run on any new dataset

Apply tangent to a DataFrame — the math node that turns garbage to NaN

Turn messy date strings into real datetime columns

The bridge from DataFrame to numpy arrays

A whole DataFrame, as one big text blob

Rows become columns, columns become rows

How often does each value show up?

Per-column variance, because std alone doesn't tell the story

Stack Two DataFrames Into One Without Touching a Terminal

Cut a DataFrame in Half — Your Train/Test Split, in Two Nodes

Selecting Rows From a MultiIndex DataFrame, Sans Panic

A Date to a Filename-Ready String, Formatted Your Way

Turn a Number Into Text — Because Some Nodes Only Speak Strings

An Integer That Needs to Be Text? One Node, No Rounding Drama

Build a One-Key Dict From a Widget — Feed It Straight Into Pandas

A Key Plus an Integer Becomes a Dict — That's the Whole Trick

Stuff a Text Value Into a Dict — Multiline Welcome

See What's Actually in That List — and Render It as Text

30:00' Into a Real Datetime Object

Type a Python Dict Literal, Get a Real Dict, Feed Pandas

Parse a Number Out of a Text Field — With All the Float Quirks

Pull an Integer Out of Text — No Float Traps, Just Strictness

Type a Python List Literal, Get a Real List — Round Trip Included

Bar Charts Straight Out of Your DataFrame — This Is the Workhorse

Bar Chart a Tensor Directly — Skip the DataFrame Altogether

Distributions by Category, in One Node — Median, Quartiles, Outliers

This Heatmap Is Really a Correlation Matrix

One Column, One Histogram — Pixel Distributions and Everything Else

Histogram a tensor directly — no DataFrame detour

Line charts in ComfyUI — no Python, no leaving the graph

Line plots straight from a tensor, when you're already in tensor land

See every column against every other column at once

Spot a correlation without leaving ComfyUI

Scatter a tensor, skip the DataFrame