Rufus
Rufus
#done/process moved to ../projects/datahub.io-design-sprint-2023
Product Options
- "Hub" (or DataHub for GitHub) Connect github (+ storage) ⟹ presentable and queryable data ⟹ share your data
- Catalog: connect data sources ⟹ catalog ⟹ find your data
- Pages: connect docs and data ⟹ data driven website ⟹ share my insights
KISS version of DataHub
- (1) Connect github repo ⟹ get a web page at datahub.io/@org/reponame
- (1a) put url of a public github ⟹ get a web page at e.g. datahub.io/viewer/@org/reponame
- This avoids even having to do login and tests the showcase logic
- (1a) put url of a public github ⟹ get a web page at e.g. datahub.io/viewer/@org/reponame
What does showcase page have?
What does repo need to have?
- datapackage.json (or datapackage.yaml?)
- README
Discussion
- why do we need datapackage.json? could we get away with no datapackage.json?
- needed for external data e.g. data in s3 etc
- needed for schemas for files (if you have them)
- needed for storing views config if persisted
- why not have it? ✅2023-02-20 one less thing for user to add. can start with repos without one
- Do we have a data API
- Yes ⟹ need to trigger the loading. Benefit is higher quality and simpler data explorer frontend
- No ⟹ faster loading. But have to build querying in frontend.
- If a data API how do we deal with build time? e.g. do we show a spinner for data views. Do we show a simple preview of the data (e.g. first 5k rows) loaded directly from raw file?