Data Tidying

From messy data to analysis-ready

A dedicated workspace for data preparation. Audit issues, apply transforms, and chat with a tidyverse expert - all while your original dataset stays untouched.

Get Started Free

Safety First

Never mutate the original

When you select a dataset in the Tidy tab, AiR creates a working copy (e.g., mtcars_tidy). Every transformation, imputation, and chat-suggested operation runs on the copy only. Your original data is always intact.

mtcars

Original (read-only)

mtcars_tidy

Working copy (all edits here)

Audit

Every issue surfaced automatically

When you open the Tidy tab, AiR runs a structured audit on your data - checking for missing values, outliers, type issues, skewness, duplicates, and constant columns. Each issue comes with one-click fix actions.

Outliers

price: 1,951 outliers detected using IQR method

WinsorizeRemoveFlag

Missing Data

income: 2,847 NAs (24.1%)

Impute MeanImpute MedianDrop Rows

Skewness

price, carat: right-skewed (skew > 2.0)

LogSqrt

Type Issues

cut, color, clarity: character should be factor

To Factor

Duplicates

12 exact duplicate rows found

Remove Duplicates

Each quick-action button applies a standard, deterministic R transformation - no AI cost, instant execution, fully reproducible.

Transforms

Build a reproducible transformation pipeline

Every transformation is logged in a visual timeline with before/after metrics. Undo any step and AiR replays the remaining transforms automatically. When you're done, export everything as a clean .R script.

21 built-in transforms

Missing Values

Impute mean, median, mode; drop NA rows; drop column; flag missingness

Outliers

Winsorize (IQR capping), remove outlier rows, flag as indicator column

Transformations

Log (log1p), square root, z-score, min-max scaling

Type Conversion

To factor, to numeric, to date, to logical

String Cleaning

Trim whitespace, standardize case, remove empty strings

Deduplication

Remove exact duplicate rows

Transform timeline

Winsorize price

Outliers: 1,951 → 0

applied

Log transform carat

Skew: 3.2 → 0.4

applied

Convert cut to factor

chr → fct (5 levels)

applied

Impute median for depth

NAs: 12 → 0

applied

Click Undo on any transform to remove it. AiR replays the remaining steps and validates each one.

Expert Chat

A tidyverse expert at your fingertips

The built-in chat understands your dataset and its current state. Ask about joins, pivots, reshaping, feature engineering, or any tidyverse operation. When it suggests code, click Apply to execute it immediately on your working copy.

Tidyverse ExpertWorking on: diamonds_tidy

Pivot this from wide to long format, keeping carat and price as id columns

Here's a pivot to long format:

diamonds_tidy <- diamonds_tidy %>%
  pivot_longer(
    cols = -c(carat, price),
    names_to = "variable",
    values_to = "value"
  )

ApplyRuns on diamonds_tidy only

Joins & merges

Pivot & reshape

Feature engineering

Pipeline

From raw data to model-ready in one flow

The Tidy tab sits between Data and Model in AiR's workflow. Explore your data, clean and transform it, then hand it off to Model Builder - all in one seamless pipeline.

Chat

Ask anything

Data

Explore

Tidy

Clean & transform

Model

Build & evaluate

Usage

Track costs

Use in Model Builder

One click to switch to the Model tab with your tidied dataset pre-selected.

Save Dataset

Export your tidied data as CSV or RDS for use outside AiR.

Export Script

Generate a documented .R script with every transform, ready to reproduce your pipeline.

Clean data, better models

Start tidying your data in minutes. No setup, no packages to install.

Get Started Free Back to Home