MINIMUM VIABLE PRODUCT

[jeanpaulrsoucy]

This issue should be used for tracking and discussing what a "minimum viable product" of the FAIR COVID-19 Data for 🇨🇦 project would look like and what tasks must be accomplished.

What would a minimum viable product look like?

A minimum viable product (MVP) for this project would likely be a functional front-end (#8) supported by an accessible and documented API (#9) for a carefully curated set of the most important datasets transformed and available in FAIR format (#10, #15).

[colliand]

We might be able to leverage available tech to get to a desired future state faster. Are there other open source front ends we should investigate?

FRDR

The FRDR resource includes these results when I search for covid-19.

Does FRDR achieve our FAIR goals? Clicking through the links in the search results above usually leads to a web page providing an overview of a data source. Some of these pages include visualization and links for downloading data. This shows some success on F and A but more work needs to be done on I and R.

Some of the data assembled and transformed by CCODWG goes deeper on I and R but is perhaps behind on F and A.

UBC Open Collections

Another front end example I like is UBC's Open Collections. Data in that resource can be explored using a web GUI that feels like an offering from a library, similar to the FRDR interface. However, Open Collections also offers a documented API with code samples.

Statistics Canada

Statistics Canada offers a front end for finding and accessing open data. Here are the results of searching that resource for covid-19. Clicking through those links usually leads to a nicely rendered table of the data with links for downloading. Exploration and discovery capacity would be improved if the resource made it more obvious about how to access the data programmatically via API.

Pangeo Gallery

The Pangeo Gallery is a resource that provides examples of code+data+visualization by leveraging ephemeral launches of customized toolchains and code with API calls using MyBinder. The gallery emphasizes I and R more than F and A*.

[jeanpaulrsoucy]

Let's not count out Datasette, although SQLite obviously has its limitations as a format for serving all of the kinds of files the Archive has to offer.

UBC Library seems to have some open code related to Open Collections website.

[colliand]

What should the MVP for the future historian studying the pandemic look like? Should we engage with a digital historian to provide guidance on how to ensure the data flowing into https://whathappened.coronavirus.icu/ is organized for future investigators? This usage scenario appears to me to provide some unification of the CANMOD Digitization and CCODWG projects.

[jeanpaulrsoucy]

@colliand geodisy looks like a fantastic starting point for our platform. It's possible we will need to extend the platform for our own needs—meaning we could give back to the original project. Even better, the prototype for the platform is hosted by FRDR: https://geo.frdr-dfdr.ca/.