First, holy cow! What an awesome resource Ryan has put together!!
Shaping data
- It's optimal to have one column representing one field (layout is vertical not horizontal)
- One recommendation for time: if it's presented as 1Q16, add a column to data for the actual date (3/31/16)
- Measure is quantitative and dependent (i.e Sales by itself just gives you a total but a dimension will give it context (sales by region)
- General rule of thumb: measure is a number, dimension is what to slice & dice by
- A good question to ask when you have numbers as data items, does it make sense to sum it up? If yes, then measure. If no, then it's likely a dimension.
- When null is the first value, Tableau may misclassify because the first value is a string.
- Discrete is blue, continuous is green pill (Emma Whyte would smack me for using the incorrect term of pill).
- Discrete fields draw headers (which can be sorted) while continuous draw axes (cannot be sorted).
- The detail marks card allows you add a dimension which can take a view from a high level of aggregation to be more granular (defined by that dimension).
- It's not the only way but another way that may help the viz process flow better.