Update on column mappings - reconciliation dialog redesign

Hi everyone, wanted to share with the forum an update on the redesign of the second stage of reconciliation: Column mappings.

The redesign breaks column mappings into stepper process of 3 steps ,where users can easily differentiate between matching as name and matching as property value and also refining with other columns.
It also simplifies reconciliation language as new users have mentioned struggling with the technicalities of reconciliation language.

You can find the conversation surrounding this issue, proposal and discussion here Propose redesign for the whole reconciliation dialog · Issue #6053 · OpenRefine/OpenRefine · GitHub

And a here's the Figma file https://www.figma.com/file/QWESrOCcNY8gDcmtByY2UO/Reconciliation-Dialog-Redesign?type=design&node-id=811-1475&mode=design&t=Hc7urYpZ2KSEdO55-4

Step 1: Matching values

Step 2: Filter results by type:

Step 3: Refine results with other columns

Step 3.1

Step: 3.2

Start reconciling

Please feel free to share feedback.


First of all: thank you for taking the effort to redesign, prototype and share.

When I teach OpenRefine I notice, that a lot of people are learning and using Reconciliation in an iterative/explorative manner. Meaning they play around with options, check the results and then tweak the options. They also tend to document and compare settings using screenshots.

But having a multi step workflow you tend to forget, what settings you chose in the steps before.
It is also quite complex to document each step instead of having an overview with all the options.

So my suggestion is to somehow show the chosen settings/options of previous steps.
But this is just an observation and an idea, not sure whether this would fit into the existing design(s).

Also, the new column mapping is going to be great when you work with projects that contain a lot of columns!

@wetneb Should we change the wording of

Automatically match results with high confidence

to something like

Automatically match results with > 75% confidence?

I worry about just saying "high" without presenting what that is dynamically behind the scenes.
But maybe "high" is simple enough and we could have a tooltip (?) icon after that sentence to hover over and explain?
Maybe @zoecooper might have some ideas here also.

@Lydiaofficial I kinda agree with @b2m where we should at least present a summary at last stepper page of only the chosen settings. This is more work overall, I know, but I also would like to see it and would use it also (being able to screenshot it, as well as grab the summary quickly as copy/paste text. It's because stepper page 1 and 2 are very dynamic for me sometimes with trial and error for Propeties and Types, as well the getting a copy of the full property mapping as text I can paste into a document of my workflow is a big bonus.