Data Is Something That People Make.

Data Can Be Rectangular (Tabular).

Data Can be Structured (Relational).

Data has Rows (Observations) and Columns (Features).

One Row for Each Observation.

One Column for Each Type of Information.

One Value in Every Cell.

Avoid Visual “Data” (i.e. colors in a spreadsheet).

Use Good Null Values.

Save Data in Plain Text Files (CSV or TSV).

Data Can Be Quantitative (Numerical).

• Discrete (integers or whole numbers)
• Continuous (any value, including decimals)

Data Can Be Qualitative (Categorical).

• Ordinal (categories have an order or hierarchy)
• Nominal (categories don’t have an order)

xkcd.com/2054

Types of Metadata

• Descriptive metadata can help you find and understand a dataset or resource. (e.g. titles and authors of books)
• Administrative metadata can help you manage a resource and tell you about how it was created. (e.g. publishers of books, or ebook file formats)
• Structural metadata can help you understand the different parts of the dataset and their relationship. (e.g. column labels, or tables of contents in books)

Metadata has standards for interoperability.

Dublin Core – 13 features common to digital data

DDI: Data Documentation Initiative – social, behavioral, and economic sciences

MARC: MAchine Readable Cataloging – libraries, books & media

xkcd.com/2239