tidy data for social science surveys

[jeanetteclark]

Need to write a lesson on how to create tidy data structures out of survey data

key points to hit:

• entities, observations, variables (survey population, individual, question response)
• consistent coding of variables
• open formats

our example dataset might feature:

• excel format with tabs?
• inconsistent coding variables
• other?

@mbjones would like your input here

[jeanetteclark]

we could also use two tables to show how to separate sensitive and non-sensitive data with participantId as a key

[jeanetteclark]

Data modeling lesson

• 10 simple rules
• entities, observations, variables
• joins and entity relationship diagrams between two tables of fake data(perhaps survey answers and community characteristics, with community being the join variable)
• group exercise, design tables based on a survey instrument

Data tidying lesson

• cleaning up text based issues (stringr) in fake dataset created for first lesson
• miscapitalization, stray whitespaces
• joins
• group_by, summarize, filter
• separate?

Text analysis lesson

• analyzing the fake dataset free text responses