Record statistics snapshots by year of publication

[anaelizabethenriquez]

Add granularity to the statistics snapshots based on year of publication. In addition to capturing the "Total article count" and "Open access article count" (scoped to pubs subject to the OA workflow, per #156), break that down by year of publication for at least each year from 2019 to present. (Ideally, we would go back a bit further, but if that makes things difficult, 2019 is adequate.)

So, we would have the following data snapshotted each week: "Total article count", "Open access article count", "% open access" 2019 article count, 2019 OA article count, 2019 % open access 2020 article count, 2020 OA article count, 2020 % open access etc.

The idea here is that we'll be better able to see the impact of our work if we isolate each year of publications. Right now the % OA changes very slowly because the denominator is big.

[anaelizabethenriquez]

Related to #167 and #274, which we should probably update and consider working on at the same time.