DocTransform
DocsGuidesData Profiling

Data Profiling & Insights

Understand your data before you change it.

See what's really in your file, before you touch it.

Data Profiling infers each column's type from a random sample of 1,000 rows, then profiles the first 1,000 rows for headline stats. Every header shows a histogram plus null, invalid and unique counts — no click required. Around 20 types are recognized, so figures, dates, emails, currencies and more are told apart automatically.

Instant histograms

Visual distributions

See the shape of every column at a glance — where values cluster, and where they don't.

Outliers and mixed types

A value more than three standard deviations from the mean (z > 3), or a value that doesn't match its column's type, is flagged automatically in the Health Hub panel as you profile.

What each header shows

1

Uniqueness count

How many distinct values exist — useful for spotting duplicates in ID columns.

2

Null analysis

The exact share of missing values, so you can decide whether to Drop or Fill them.

3

Type inference

The detected type — Money, Date, Text, and around 20 others — plus a flag when a value doesn't match it.

Before and after

clinicflow_visits.csv
Patient IDAgeVisit Date
H-1042342026-01-14
H-10431422026-01-15
H-1044N/A2026-01-16
  • 1 Raw import, before profiling.