Skip to main content

What is VisiData

VisiData is an interactive multitool for tabular data that runs entirely in your terminal. It reads virtually any structured data format, presents it as a navigable spreadsheet-like sheet, and lets you reshape, filter, and analyze data using only keyboard commands — no mouse, no GUI, no waiting.

Learning Focus

Use this lesson to build VisiData's core mental model: everything is a sheet, every sheet lives in a stack, and every transformation generates a derived sheet — not a destructive edit.

Tool Snapshot
  • Command: vd or visidata
  • Language: Python (GPLv3)
  • Formats: CSV, TSV, JSON, SQLite, Parquet, Excel, fixed-width, logs, URLs, and more
  • Version (stable): v3.3 (2025)
  • Config file: ~/.visidatarc
  • Help inside VisiData: Ctrl+H (man page), Alt+H (menu)

The Problem VisiData Solves​

Imagine you have a CSV file with 500,000 rows. Opening it in Excel takes 45 seconds and may freeze your machine. With VisiData, the same file opens instantly — because VisiData loads data lazily (only what's visible on screen).

Opening a large CSV in a GUI spreadsheet:

$ open 500k_rows.csv        → 45 seconds loading
  filter one column          → application freezes
  format change              → out-of-memory warning

Opening the same file in VisiData:

$ vd 500k_rows.csv          → loaded instantly (lazy loading)
  Shift+F                    → frequency table in <1 second
  q                          → back to source sheet instantly

How VisiData Is Structured​

Every file you open is a sheet. Every transformation (frequency table, pivot, summary) pushes a new derived sheet on top. Press q to pop back. Nothing destroys the original.

Source Sheet ──Shift+F──▶ Frequency Table ──q──▶ back
             ──Shift+W──▶ Pivot Table     ──q──▶ back
             ──Shift+I──▶ Describe Sheet  ──q──▶ back

For the full sheet stack model, see Sheet Stack and Prefix Keys.

What VisiData Looks Like​

Suppose you open a simple CSV file of employees:

name,city,department,salary
Alice,Jakarta,Engineering,90000
Bob,Bandung,Marketing,70000
Charlie,Jakarta,Engineering,85000
Dina,Surabaya,HR,60000

Inside VisiData it looks like this:

┌──────────────────────────────────────────────────────────┐
│  name        │  city        │  department  │  salary      │  ← Header row
│──────────────┼──────────────┼──────────────┼──────────────│
│  Alice       │  Jakarta     │  Engineering │  90000       │  ← Cursor row (highlighted)
│  Bob         │  Bandung     │  Marketing   │  70000       │
│  Charlie     │  Jakarta     │  Engineering │  85000       │
│  Dina        │  Surabaya    │  HR          │  60000       │
│──────────────────────────────────────────────────────────│
│  employees.csv   │  4 rows   │  name                      │  ← Status bar
└──────────────────────────────────────────────────────────┘

Key interface elements:

  • Header row — column names (press ^ to rename any column)
  • Cursor row — the currently highlighted row your cursor is on
  • Status bar — shows sheet name, total row count, current column name

Why Operators Use VisiData​

NeedVisiData capabilityPractical effect
Fast large-file inspectionLazy loading, 1M+ rowsNo GUI freeze
Filter/slice dataRow selection, regex, Python exprReplace grep+awk pipelines
Data type casting~#%$@ type keysInstant int/float/date coercion
Frequency analysisShift+F frequency tableInstant histogram
Data exportCtrl+S to any formatCSV → JSON → SQLite in seconds
RepeatabilityCommandLog replayDocumented, reproducible transforms
Custom logicPython expressions in columnsFull pandas-like power in-terminal

VisiData vs Alternatives​

FeatureVisiDatacsvkitMillerpandas
Interactive TUI✅❌❌❌
1M+ rows✅✅✅✅ (with RAM)
Multi-format✅ 40+LimitedTSV/CSV/JSON✅
Python expressions✅❌❌✅
In-terminal graphs✅❌❌❌
Learning curveMediumLowMediumHigh
tip

VisiData is the best choice when you need interactive, visual exploration of data without leaving the terminal.

Common Beginner Mistakes​

MistakeWhat happensFix
Pressing q to quit VisiDataOnly closes current sheetUse gq to quit all
Editing cells unintentionallyPressing e enters edit modePress Ctrl+C to cancel
Losing the source sheetNavigating derived sheetsPress Shift+S to see all sheets
Confusing column typesNumbers treated as stringsSet type with # (int) or % (float)

Basic Commands Overview​

# Open a file
vd ~/github/practice-folder/visidata/01-loading/01-employees.csv
vd ~/github/practice-folder/visidata/05-logs/01-access.log
vd ~/github/practice-folder/visidata/08-sysadmin/site.sqlite

# Inside VisiData:
# q         quit current sheet
# gq        quit all sheets (exit VisiData)
# Ctrl+H    open man page / full reference
# Alt+H     open help menu
# Shift+S   open Sheets Sheet (see all open sheets)

Hands-On Practice​

# Create the sample file
cat > ~/github/practice-folder/visidata/01-loading/01-employees.csv << 'EOF'
name,city,department,salary
Alice,Jakarta,Engineering,90000
Bob,Bandung,Marketing,70000
Charlie,Jakarta,Engineering,85000
Dina,Surabaya,HR,60000
EOF

# Open it
vd ~/github/practice-folder/visidata/01-loading/01-employees.csv

# Try these inside VisiData:
# hjkl or arrow keys — move around
# Shift+F            — frequency table on current column (try on 'department')
# q                  — return to source sheet
# gq                 — quit VisiData

What's Next​