How to Get Started with VisiData

Updated

Install VisiData, then open a file from your terminal with vd input.csv. You'll land in an interactive view of the data where single keystrokes give you summaries, plots, and filters. This guide covers the first steps: installing, opening a file, using the built-in tutorial and quick reference, trying the key exploration features, and saving or replaying a session.

Before you start

VisiData is a command-line tool, so you need a terminal and a working Python environment. The site describes it as "a lightweight utility which can handle millions of rows with ease," combining "the clarity of a spreadsheet, the efficiency of the terminal, and the power of Python."

You don't need to know Python to start, but Python knowledge becomes useful later for per-row computation and custom commands.

Step 1: Install VisiData

The site's Install link is the canonical starting point. Follow the instructions there for your platform rather than copying commands from elsewhere, since installation methods differ by operating system.

Expected result: the vd command is available in your terminal. You can confirm this by running vd --version; the current release listed on the site is v3.4 (Jun 30, 2026), with v3.3, v3.2, and v3.1 as earlier releases.

Step 2: Open your first file

From the directory containing your data, run:

vd input.csv

The site's own example uses exactly this form. VisiData opens the file as a scrollable sheet. The same command works for the many formats the site lists, including:

  • CSV, TSV, and fixed-width text
  • Excel (.xls/.xlsx)
  • JSON and JSON-Lines
  • HTML, YAML, Markdown, and org-mode
  • SQLite databases (.db), HDF5, NumPy (.npy/.npz), and Pandas DataFrames
  • SAS/SPSS/Stata files (.xpt/sas7bdat/sav/dta)
  • Shapefiles (.shp), packet captures (.pcap), .zip archives, and git repositories

Expected result: your data appears in the terminal, ready to navigate. The site notes it can "open 1m+ rows in the blink of an eye."

Step 3: Use the built-in tutorial and quick reference

The homepage links directly to a Tutorial and a Quick Reference Guide. Start with the tutorial if you've never used a terminal spreadsheet before; keep the quick reference open in another window while you work, since VisiData is keystroke-driven and you'll be looking up commands constantly at first.

The Documentation and "10 Ways to Use VisiData" pages are useful next stops once the basics feel familiar.

Step 4: Try the headline exploration features

Instant histogram

The site advertises an "instant histogram with one keystroke." Select a numeric column and trigger the histogram to see the distribution without writing any code. This is the fastest way to spot outliers or skewed data in a new dataset.

Scatterplots in the terminal

VisiData can draw "scatterplots in the terminal," so you can check relationships between two numeric columns without leaving the command line or exporting to a plotting library.

Python per row

For calculations the built-in commands don't cover, the site lists "compute Python for each row" as a supported workflow. This is where the Python integration pays off: you can derive new columns or transform values in place.

Expected result: after a few minutes you should be able to open a file, scan a column's distribution, and plot two columns against each other, all from the keyboard.

Step 5: Save sessions and run batch pipelines

Two features make VisiData more than a one-off viewer:

  • Save sessions for replay later. If you've set up filters, derived columns, or a particular view, you can save the session and return to it instead of redoing the work.
  • Adhoc data pipelines in batch mode. The site shows vd -b input.csv as the batch invocation. The -b flag runs VisiData non-interactively, which is how you chain it into scripts.

Expected result: repeatable exploration. Save a session when you expect to revisit a dataset; use batch mode when the same transformation needs to run on a schedule or across many files.

Common sticking points

  • Nothing happens when you type. VisiData is modal and keystroke-driven; plain typing doesn't enter text. Check the Quick Reference Guide for the command you want, and look for an on-screen hint about the current mode.
  • The file won't open. Confirm the path is correct and the format is one VisiData supports. Most common tabular formats are covered, but unusual binary formats may need conversion first.
  • You forget a command. The built-in tutorial and quick reference exist precisely for this; the site also points to a Lightning Demo and video case studies for visual walkthroughs.

Where to go next

Once you're comfortable opening files and exploring columns, work through the "10 Ways to Use VisiData" page for concrete workflows, and check the blog for updates. The site also invites support via Patreon if you find the tool useful.