Data Management Planning: Documenting & Structuring Your Data

Document:

• Data collection methods
• Context of data collection
• Variable names and description
• Algorithms used
• Transformations of data from the raw data through analysis
• Software and systems used for analysis

Use a script rather than GUI during data analysis, better for documentation and makes results easier to reproduce

Incorporate a workflow tool such as Kepler, Taverna or VisTrails

A consistent convention will help you easily identify your files and what they contain. Use abbreviated descriptive information such as

• project
• content or parameter
• location, date and/or time (yyyymmdd for easy sorting; hhmmssTZD for time)
• version number (establish numbering system for versions)

Use numbers, letters, dashes, underscores. Do not use spaces or special characters. Stay concise to be practical.

Define a data dictionary:

Use discipline specific metadata standards:

• Metadata editors - a good list from DataONE
• Metadata: Distributed Active Archive Center (Oak Ridge National Laboratory)
• Schema Library (Altova) of common industry and cross-industry standards
• Seeing Standards: A Visualization of the Metadata Universe documents 105 standards used by the cultural heritage community
• DDI: Metadata specification for the social and behavioral sciences

• Use in conjunction with a "Data Dictionary" containing information about:
  • Variable name
  • Variable types
  • Codes and Ranges
  • Missing values
• Place variable names in row 1
• Always have a unique identifier per entity
• Keep track of changes made to worksheet
• Format columns to matchthe variable type (date, numeric, text, etc.)
• Data entry guidelines:
  • Freeze column headings so they will not scroll off the screen
  • Enter string variables in a consistent case
  • Do not leave any blank rows in the spreadsheet
  • Do not include unessential text or fancy formatting in the spreadsheet
  • Get rid of formulas - copy the entire spreadsheet into a new sheet using "Values" option
  • Sort data with caution (always SAVE first) 
• Verify data using double data entry
• Save as .csv for forward compatibility and interoperability

Resources:

• Elliott, A C. (2006). Preparing data for analysis using Microsoft Excel. Journal of investigative medicine, 54(06), 334-341.