read-data-dict

name	read-data-dict
description	Read and understand a data-dict.yaml file so you can use it when working with the dataset it describes.

Read a data dictionary

Read and understand a data-dict.yaml file so you can use it when working with the dataset it describes.

Steps

1. Find the data dictionary. Look for a file called data-dict.yaml in the project root or the same directory as the dataset. If there isn't one, tell the user and stop.

2. Read the file. Parse the YAML and familiarise yourself with its three top-level sections:

   • tables -- the tables, their columns, types, constraints, and descriptions.
   • relationships -- how the tables join together, including cardinality and any column-name conflicts.
   • glossary -- domain-specific terms and their definitions.

3. Internalise the glossary first. The glossary defines the vocabulary used throughout the rest of the file. Read it before interpreting column descriptions so that domain terms are understood in context.

4. Use the dictionary as context. When answering questions about the data, writing queries, or generating analysis code:

   • Respect column types and measures (e.g. don't average an id column).
   • Honour constraints (e.g. primary keys are unique and non-null).
   • Use relationships to determine correct joins and carefully watch for conflicts.
   • Use source to determine how to access the data.