Data Documentation

Arguably, the most important element of working with any dataset (GIS or not).

Federal spatial datasets metadata guidelines from FGDC

A nice summary of the FGDC metadata standard

See also the FGDC Metadata workbook

Google is using crowdsourcing to update their Google Maps data...

These guidelines may well be overkill for many projects, but that doesn't mean that documentation isn't necessary. It is critical to provide some degree of data documentation for every dataset. Depending on the project, you may or may not want to produce metadata that meets federal or international standards for spatial data documentation. It depends on the nature of the project and why you are documenting. But no matter what, you should always produce some metadata.

ArcGIS contains a very nice set of tools that allow you to produce FGDC compliant metadata. These tools are available through the ArcCatalog interface by selecting any dataset and clicking on the Metadata tab. By default this tab will display the metadata for the selected dataset (if such exists). There are tools in the ArcCatalog interface that allow you to 1) create/edit metadata and 2) import/export metadata.

Metadata in ArcGIS is stored in a series of XML files (XML is a markup language just like HTML). These XML elements include properties and documentation. Properties are things that ArcGIS can update itself (huge timesaver) such as the coordinate system of a dataset, the domain of a dataset, the number of features and so on. Documentation is what you provide to describe the characteristics of the dataset.

Here is a set of key things to document:

1) source of dataset - where data came from
2) accuracy of dataset - spatial and attribute
3) lineage of dataset - processing steps performed on the dataset
4) attributes of dataset - description of attributes
5) time period information
6) contact information - citation, contact, author, metadata author
7) spatial reference information (coordinate system, map projection, datum)
8) format
9) spatial domain/extent
10) scale (if not implicit) - resolution

Beyond documentation of spatial datasets using the FGDC standard, it is important to document files and directories both for yourself and for others who may use those files/directories in the future. This can be done using the following files in the folders/directories in which you are working:

README
CONTENTS
README.<covname>
README.<processing>

Alternatively, files along these lines can be constructed as HTML files and then viewed in ArcCatalog.