"Caliperize" Hypothesis interactions

From what I'm learning, being able to deliver annotation activity data in a readily usable format to schools would be both 1) a major selling point when it comes to initial enterprise sales and 2) a premium add on to a base cost for use of the LMS app product.

It's clear to me that that format is Caliper Analytics, a data standard produced by the same IMS group that developed the LTI standard. They already have an ["annotation profile"](https://www.imsglobal.org/caliper-11-metric-profiles#annotations:-AC_ZjuSEemlZovu36inXQ) among other profiles that we could produce data for. We are in touch with one of the main architects of Caliper who has asked for our feedback about the annotation profile. Attached is a nice intro deck to Caliper that he has shared with me.

[arwhyte-caliper-hypothesis-20190228.pdf](https://github.com/hypothesis/h/files/3118936/arwhyte-caliper-hypothesis-20190228.pdf)

In the medium term, I recommend we produce Caliper data packages from annotation activity specifically within the LMS app. My vague understanding is that this will involve querying our API and repackaging the data to send via Caliper API. The end goal would be sending data about institutional annotation activity with Hypothesis in Caliper format to whatever data storage service that institutions is using.

In the short term, I'm hoping that @judell can do some initial exploration of Caliper and how we might translate Hypothesis annotation activity into that format.