Skip to content

Stockertim Dictionary · Artificial intelligence · Technology

Data dictionary

Also called data dictionaries, codebook, field dictionary

In short

A document that lists each field in a data set with its meaning, its type and the values it is allowed to hold, so everyone reads the data the same way.

  • Beginner
  • Document
  • Checked

In detail

A data dictionary describes a data set field by field. For each column it records the name, a plain explanation of what the column means, the data type (text, whole number, decimal, date), the unit, the allowed values or range, whether blanks are permitted and where the data comes from.

Its purpose is shared understanding. A column called status might hold order status in one system and payment status in another; a date might be the order date or the dispatch date; an amount might include or exclude tax. Without a dictionary, two analysts can produce different answers from the same file, both convinced they are right.

A data dictionary is also one of the most useful things to give an AI assistant. Describing the columns precisely lets the assistant write formulas, queries and cleaning steps without seeing the real records, which protects personal and confidential data. A vague description such as "a sales file" produces vague and often wrong formulas.

Dictionaries need maintaining. When a field is added, renamed or redefined, the dictionary is updated with the date, so older analyses can still be understood.

Worked example

Kite & Co Accountants prepares to ask an assistant for help analysing an invoice extract.

  1. Weak prompt: "Here is our invoice spreadsheet, tell me what is overdue."
  2. The team writes a data dictionary instead: invoice_id (text, unique), client_region (North, South or East), amount_gbp (decimal, excluding VAT), due_date (date, day/month/year), status (Paid or Unpaid).
  3. Improved prompt: "Using the columns described below, write a formula that totals amount_gbp for Unpaid invoices whose due_date is before today."
  4. The assistant writes the formula without seeing any client data, and the team tests it on five known rows.

The formula works first time, and no client record left the firm.

In practice

Data teams keep dictionaries alongside databases, spreadsheets and dashboards. Analysts attach them to data requests, and good dictionaries record who owns each field and when its definition last changed.

Watch out

  • A column name is not a definition; similar names in two systems can mean different things.
  • An out-of-date dictionary misleads more than having none, so date every change.

Learn it properly

Official sources

Facts last checked on 3 November 2026.

Education, not advice: the Stockertim Dictionary explains what terms mean. It never recommends buying or selling anything.

English