Data dictionaries can be valuable as a source for understanding the types of variables under analysis and how they are measured. Here are some useful metadata items to keep in a data dictionary:
- Name of the variable: Consistency in naming conventions helps in understanding and readability. Some analysts like to use CamelCase, others like to use punctuation for an object, as in Variable.data.frame, and others will insist on only having lowercase letters.
- Measurement data: This answers questions such as, "Is the data numeric or categorical?", "How many levels are contained in each category?", and "What is the length of each variable?"
- Sources of the data: This covers, "Where did the data originally came from?"
- Transformations ...