In order to keep your research data usable to yourself and others, it is recommended to document your data: record all the information necessary to understand the content and context of the data.
At the end of the research project you are probably unable or unwilling to spend a few more days documenting your dataset(s). Moreover, by that time it is doubtful if you still remember all the decisions you made during the project regarding your data.
Therefore it is advisable to pay regular attention during the research project to the documentation of your data and to record this information in a text file stored with your data.
In your documentation you need to include information on the collection of the data, the data input, the data storage and the data processing. Lab journal or codebook also count as documentation of a dataset and are preferably stored with the data.
All the information necessary to be able to evaluate and (re)use the data over a longer period, for example five years, must be in the documentation. For example, such information as
- context and methodology of the research
- the (settings of the) apparatus or instruments used
- the relation between files in a dataset or database
- the content and structure of individual files (variables, etc.)
The above is information meant for humans, also called human readable metadata. There is still another type of documentation: assigning metadata to be read by computers and search engines.