Calculate the most common failures across collections

[jpmckinney]

Since this will only be used on occasion, and mostly to inform the development of implementation resources and the standard itself, we can take a low-fidelity approach, e.g. just running SQL queries – rather than a high-fidelity approach, e.g. with a designed results page.

[jpmckinney]

This has not come up again as a need. We already have an idea of common issues through helpdesk support to publishers.

In general, anytime there's a quality issue, we should consider whether/how to improve schema descriptions, documentation, guidance, and the boilerplate text explaining how to address failing checks.

open-contracting/pelican-frontend#53 will also help in a non-aggregated way.