Conversation
|
Following up from the Thursday meeting, I have a couple of ideas:
Happy to pick this PR further. |
|
The options @chinardankhara mentions are both useful. Not sure how multiselect plays with autocomplete, but if it filters on the fly, I don't see why we need to stop at <=5 items? We do want to support arbitrary number of selections. Removing the checkboxes is fine as long as we have some alternative. CSV upload could initially just be OpenAlex IDs (or ROR ids, or URLs with a mix of openalex and ror API calls). That should be simplest to implement (less work on our end and shows proof of concept). Having users input arbitrary institution names brings up a lot of other UI/UX issues. Even if we sanitize/fuzzy search for a better initial list, there will be subtle errors and the need for user validation creating extra steps and potential unresolvable ambiguities that will require users to input PIDs like OpenAlex or ROR ids anyway. I think for a first pass, requiring a single column CSV with OpenAlex and/or ROR URIs seems like we can at least provide results cleanly, even if the user has to work harder pre-upload. That will allow us to get the visualization and data components in place and then we can work on simplifying user input. Happy to chat pros and cons of these and other approaches further. This is a design decision which won't likely be the final one. Another consideration is the uploading a CSV of people should be similar to Orgs, and a list of Open Alex IDs eliminates the harder name disambiguation |
…elds are blank, as a first pass
|
I've added an update which allows you to upload a .csv file with OpenAlex IDS such as this: Right now, in order for it to work the other fields must be empty because it defaults to them and I've left the checkbox to add another institution As you mentioned this can be updated in the future, but this is a first update in order to allow uploading the csv to function. |




Adds the ability to search for multiple institutions