Closed groups that are also sharply dated

The same figure as talk/img/talk-closed-groups.svg, except that it is not a figure: the query below runs against the published graph in this browser, and the plot is whatever it returned. Move a threshold and the selection changes in front of you.

There is no SPARQL endpoint and no server. The bundle is fetched as a static file and parsed with rdflib under Pyodide. That is deliberate: an archived copy of this repository stays queryable without a service having to be kept alive for it.

Loading the Python runtime (~10 MB on first visit)…
the same dies recur — a hoard, a consignment, a store
the potters agree, relative to the span they describe
enough material for k to have come off its maximum

Counting starts once the graph is loaded.

The two quality axes are never multiplied into one score. Asking for both at once in a query is how they are combined: the threshold stays visible, it is stated per axis, and you can disagree with it by dragging it.

The query

Editable. The sliders rewrite the FILTER line; anything else you type stays. The plot needs ?site ?findspot ?from ?to ?midpoint ?qInterval ?qRepetition — drop one and you get the table without the chart.

.rq

Press Run.

qinterval — box fill, red to green qrepetition — bar in the left margin | midpoint