Project Catalogue

[clairerye]

User Story

As a 'researcher with a computer' I want a central point to access all data relevant for HCA research As a stakeholder I want to know how much HCA data is available and stratify it for various purposes As a 'researcher with a pipette' I want my data shared with the community

Suggested requirements to support the data release goals:

Contributors and researchers must have a central catalog showing the full breadth of the data at a project level (individual data file level is not required). Researchers and collaborators must be able to obtain access to the data, but this can be via direct download or link out to appropriate archives/store for metadata Ideally, researchers and collaborators should be able to search the catalog to identify projects of interest (for example, filtering by organ, technology, date)

Nice to have

Researchers and collaborators should be able to see a “project view” summary page giving a high level overview of the project and links to data

Current list of planned features

Needed before 'first release':

• [x] implement solr indexing and search functionality (@jacobwindsor)
• [x] change visible on page language to 'catalogue' (@jacobwindsor)
• [x] change date format to dd/mm/yyyy (@jacobwindsor)
• [x] get ebi url/domain and host site there (@jacobwindsor once domain acquired)
• [x] small ui tweaks (@gabsie @jacobwindsor)
• [x] setup google analytics - what are people searching? from where? #140
• [x] ensure all styling meets ebi guidelines - visual framework, header footer (@gabsie)
• [x] Have a way to contact i.e. send feedback, add their own project/publication (@jacobwindsor implement/@gabsie advise on design ticket #127)
• [x] Layout for the about page (@gabsie #126 )
• [x] Content for the about page (@gabsie @mshadbolt #138)
• [x] Create static about page that links from catalogue page ticket (@jacobwindsor )
• [x] decision on name - "HCA Project Catalogue"
• [x] remove wranglers/curators from contributor list (@jacobwindsor )
• [x] add more projects not in dcp, e.g. ones in ega (@mshadbolt )
• [x] get and implement any actionable feedback from John Marioni (@lauraclarke has sent initial email)
• [x] display geo accessions and links
• [x] add a count of projects
• [x] display author names nicely
• [x] get list of organs and/or organ parts per project to enable search by organ (JM high priority)

Post-release ideas

• add dbGaP links
• add other supplementary links
• Ask for feedback from Sarah and/or Kerstin
• get list of annotated, harmonised cell types and enable search by cell type (JM high priority)
• have links to contributor matrices
• get list of assay type/technologies per project to enable search by tech
• enable ontology powered searching for the above 3 ontology based fields
• create endpoint in core to serve the required info rather than scraping
• hover over author names to show role/institute
• Distinguish 'official' HCA publications
• Distinguish between primary data pubs, integration/re-analysis papers
• Filter on grant funding org/grant number
• Filter on projects part of seed network or other consortium, e.g. WSSS, HUTER, eSPACE etc.
• show cell numbers

The current repo is: https://github.com/ebi-ait/projects-index The current link is: https://ebi-ait.github.io/projects-index/#

drive with plan and notes: https://drive.google.com/drive/folders/16kKxsrTaZSgmCP-6Lpz3drcIyjzJAUI3?usp=sharing

[mshadbolt]

proposed potential URLs:

http://ebi.ac.uk/human-cell-atlas/projects 
http://ebi.ac.uk/human-cell-atlas/project-catalog
http://ebi.ac.uk/human-cell-atlas/project-index

http://ebi.ac.uk/hca/projects - Gabby
http://ebi.ac.uk/hca/project-catalog
http://ebi.ac.uk/hca/project-index

[mshadbolt]

contacted web prod to get ebi.ac.uk/humancellatlas domain - ticket: www-prod #472626

[mshadbolt]

@jacobwindsor I added 'supplementary_links' to the data.json, I am not really sure the best way to display them, I guess we would just need to use the link text as the display text because it can contain any kind of link

[mshadbolt]

I have added all HCA pubs and various others that seem appropriate from the tracker.

I will do another review to see if there are any more in ingest that aren't appearing.

[mshadbolt]

General status update:

• web prod are working on hosting the site at the url we chose, should be live in the next day or so
• solr implementation is still chugging along, @jacobwindsor doing a lot of reading and learning to understand how to set up. Will be set up on ingest infrastructure and will need edits to the API in order to make functional
• We decided not to implement pagination for now, this will need to be implemented when we are querying the ingest API rather than loading from a json
• @gabsie & @jacobwindsor to set up a meeting to go through small ui tweaks together
• google analytics - we need to decide exactly WHAT we want to track, but we are bit blocked on full implementation until we have the final URL. @gabsie to let @jacobwindsor know the account details for the ingest GA account when he starts working on that. Can use some of the basic setup already implemented in ingest. (detailed discussion on #140)
• We decided to avoid complication of having extra pages we could have a short description at the top of the page that is expandable, gabs and I to work on text then provide to @jacobwindsor (#126)
• @gabsie to provide some input on the metadata properties, the embl ones seem a bit weird and not overly relevant but Gabs to review and ensure we are providing the important ones for google search etc (#139 )

If I forgot anything in the above please add !

[tburdett]

Great summary, thanks. I can add a view on the goal of what we want to track.

The goal (from the UX research plan) says our aim is to:

Provide scientists that are trying to find all HCA data assets with a trustable and comprehensive list of available data from all projects and publications that contribute to building the Human Cell Atlas

For me, this means I would expect users to follow a "data discovery" journey. For example, a user might search for some data by a criteria they expect results for (e.g. tissue type), review the results, and then follow links out to archives in order to download datasets that interest them. We should be able to demonstrate this by checking what people search for and detecting when they click through on an archive link. We will be doing a good job if users a) get results corresponding to the thing they searched for and b) follow links to archives for a high number of search results returned.

This is just my thoughts, though - totally open to input if there are other workflows we expect that I am missing.