The same is true for Tissue metadata. It is expected that filtering will provide the most value for consumers searching for datasets.
All Collections (consumers) or My Collections (curators) views
Currently, these views are implicitly sorted by the most recent collection. Metadata columns are limited to:
Collection (name)
Tissue
Assay
Disease
Organism
Cell Count
Consumers can sort All Collections by:
Recency using the Collection (name) column. This would depend on the internal created_at and updated_at fields in the portal database or their public replacements.
Cell Count
It is possible to sort by Collection (name) after stripping articles like 'A' from collection names such as 'A single-cell transcriptional roadmap ...' but the value is questionable based on the variations in naming by publishers.
Recency using the Dataset (name) column. This would depend on the internal created_at and updated_at fields in the portal database or their public replacements.
Cell Count
It is possible to sort by Dataset (name) after stripping articles like 'A' from dataset names, but the value is questionable based on the variations in naming by publishers.
Story
While browsing collections on the portal, data consumers may sort a limited set of metadata columns to find datasets.
Note: Sorting while filtering collections will be considered separately.
UX Design TBD Figma
Product Design
Why not allow consumers to sort all metadata columns?
Due to categorical values (which includes
organism
in schema 2.0.0), sorting in all columns might be less useful for some views.For example, 'normal' disease states will not sort in proximity because multiple collections have 'diseases+normal' such as:
The same is true for Tissue metadata. It is expected that filtering will provide the most value for consumers searching for datasets.
All Collections (consumers) or My Collections (curators) views
Currently, these views are implicitly sorted by the most recent collection. Metadata columns are limited to:
Consumers can sort All Collections by:
created_at
andupdated_at
fields in the portal database or their public replacements.It is possible to sort by Collection (name) after stripping articles like 'A' from collection names such as 'A single-cell transcriptional roadmap ...' but the value is questionable based on the variations in naming by publishers.
A Collection View
Currently in A Collection, datasets are implicitly sorted by descending order of cell counts as a mitigation for Curators want to order the datasets in a collection based on their relevance. In addition, 50% of the current collections contain only one dataset. Metadata columns are limited to:
Consumers can sort A Collection by:
created_at
andupdated_at
fields in the portal database or their public replacements.It is possible to sort by Dataset (name) after stripping articles like 'A' from dataset names, but the value is questionable based on the variations in naming by publishers.