Re: User studies: what are the pain points in reconciliation workflows?

I have made a first attempt at solliciting feedback in the VIVO
community, which hosts a number of OpenRefine reconciliation endpoints:

https://groups.google.com/forum/#!topic/vivo-community/QoSU5Pwf2YY

I have also created a dedicated tag to group existing OpenRefine issues
which require some work on the API itself:

https://github.com/OpenRefine/OpenRefine/issues?q=is%3Aissue+is%3Aopen+label%3A%22reconciliation+API+design%22

Feel free to join the effort!

Antonin

On 7/20/19 7:15 PM, Antonin Delpeuch wrote:
> Hi all,
>
> We have already identified some possible improvements to reconciliation
> in the past discussions (such as global properties or scores broken down
> in multiple features). I think it would be good to gather this sort of
> feedback more systematically, by asking and users and service providers
> to let us know what they would like to change.
>
> I realise it is not always easy to determine the scope of these "pain
> points", especially as an end user. Should a particular problem be fixed
> in OpenRefine itself, in the definition of the API, in a particular
> reconciliation service, or even in the data exposed by this
> reconciliation service? Users are not always aware of the boundaries of
> responsibilities behind the scenes.
>
> For instance, we have recently received these comments in the OpenRefine
> bug tracker:
>
> https://github.com/OpenRefine/OpenRefine/issues/2083
>
> I think this is extremely valuable and it would be great to get other
> reports in the same spirit, even if not all the comments are relevant
> for our work on the API. We can then triage them to the appropriate
> project and identify the ones that are in scope for us.
>
> I am not sure how to conduct such a campaign for feedback though. We
> could set up a web survey, but that would probably constrain the sort of
> feedback we get to the narrow frame that we pre-define (and we would
> contribute to the ambiant "survey fatigue"). I think it would be
> feasible to just request some free text feedback on various mailing
> lists and then analyze and triage the results manually, given that I am
> not anticipating hundreds of replies.
>
> We can also analyze a lot of feedback that is already public: for
> instance, I can think of reconciliation-related issues in the OpenRefine
> bug tracker, on the OpenRefine mailing list and potentially on other
> platforms?
>
> What do you think?
>
> Cheers,
>
> Antonin
>
>
>
>
>

Received on Monday, 22 July 2019 14:52:57 UTC