search

data2health / scits-platform

Science of translational science research platform
3 stars 0 forks source link

Add facet types to search #25

Open mellybelly opened 5 years ago

mellybelly commented 5 years ago

Currently its not clear what the types of things are that you can search for are. I keep clicking on sources at the bottom thinking they will take me to searching only those types/sources, but they are linkouts.

@jmcmurry could provide some icons for entity types: edu resources, datasets, clinical trials, clinical instrument, redcap survey, grants, etc.

kristiholmes commented 5 years ago

This is another reason why we need to consider the research artifact types more carefully. A hierarchy can help drive a useful faceted search. We have this as a deliverable for the repository - these should align.

mellybelly commented 5 years ago

Great idea Kristi!

We have the very many types from the requirements analysis/workshops, but would be good to standardize entity types across all the CD2H projects. From the repo landscape analysis, do we have a sense for what the upper level types are?

I think everyone (including me) is convinced now that a hierarchy of artifacts is needed for faceting and querying. Question is, what is the best way to create and maintain it, and not duplicate other efforts.

mellybelly commented 5 years ago

@newgene @eichmann also have a look at https://biolink.github.io/biolink-model/, we can make enhancements there, also should be aligning this with other key resources (wikidata, schema.org, etc)

kristiholmes commented 5 years ago

Please include @jpastva in all discussions. Please also include @saragon02 since this ties into her metadata work with @newgene.

I don't think we had any epiphanies from the workshops WRT object types, @mellybelly. it is the same stuff that shows up elsewhere.

Northwestern started this process by taking several existing schema already used in RIS systems and mapping them against each other to identify common elements and gaps. everything was mapped to the NISO research objects since that list of terms/categories was the result of a formal process that went through rigorous peer review.

We need to have some sort of conversation IRL or on slack to get on the same page about how to move this forward. Too many suggestions of potential next steps is challenging without some discussion to put the suggestions in context. I am sure we all have good examples, but a little discussion about it will save time in the long run.

How about if I schedule a 30-minute call this week for key people or their proxies?

mellybelly commented 5 years ago

happy to have a discussion but in the meantime, I suggest that we keep it simple at first: educational resources, datasets, clinical trials, software tools, expertise, publications, grants. we should do some requirements/user testing before we decide how much of the very many entity types we'd want to show to a user.

mellybelly commented 5 years ago

(sorry re-read- I meant just for addressing this ticket - lets find another place to put more extensive requirements)