As part of an existing data mapping workflow, we needed a way for users to be able to load unique data points, that don't map directly to an existing field, so they can use these when segmenting.
DataVision
Custom Field Creation
Key Goals & Requirements
Create a way for users to add custom fields when loading data
System needs to know how to store the data
Solution must be easily understandable
Initial Discovery
I needed to work closely with engineers to understand how they store various types of data as well as what options are available for how that data could present in a filter. While data experts and engineers may have a full understanding of the differences between the various data types, I wanted to create a really simple workflow that didn't rely heavily on a users understanding of the differences between an integer and a string.
What are some examples of common data points and how are they stored?
How will a user understand what data type to assign to the field?
Are we able to auto suggest based on example values in their file?
How will they know what filter type to select?
Rather than presenting a selection of technical data types, and relying on the user to know what to choose, it made more sense to present selections based on how the user is thinking, and focus more on the value itself. So technically something is a "string", but to the user it's "text".
Numbers can get particularly confusing, should it be an integer or is this number an ID and should therefore be stored as a string.
If a user were to choose "Numbers" they would then see an additional screen where they can further clarify and the backend then determines both how the field is stored and what filter option is appropriate. If the user chooses "Numeric Codes" they have the option to decide how this presents as a filter. Clicking the eye icon gives the user an example preview.
