Improve look and feel for the HTML about dataspice

[cboettig]

I started playing around with what a "card view" might look like on the plane, quick sketch here (in Hugo): https://cboettig.github.io/dataspice-web/ex/

Clearly needs a lot of tweaking but I think with the menu (click on the icons -- maybe that's too subtle) it may be possible to pack quite a bit of metadata into a small space. hope to find a bit more time to fiddle with that at some point. My thought here is mostly that a small card format could be embedded on other sites, e.g. a "Data" page on a lab group website or as a margin item on an analysis .Rmd.

I'm sure there's no good one-size-fits-all visualization anyway (I think the real utility is mostly in generating the metadata easily more than styling it), but some good default styling is no doubt helpful in motivating adoption!

One thought with the cards: in addition to adding a map, I do wonder if it would be nice to support the ability to add a "feature image" for the Dataset. Google seems to like doing this kind of thing, image is a valid attribute for a Dataset and just about everything else...

[khondula]

looks great @cboettig!

maybe if there is no image, the default thing in that place on the card could represent the coverage (map or timeline).

For temporal coverage, would something like timevis be too much?

[amoeba]

Cool, looks pretty slick. I like the idea of an embeddable card as an option. I am also really interested in (1) a full similar to what we have now but just nicer and (2) @maelle 's idea to use Hugo to collect many datasets from many creators (effectively making an easily bootstrapped, static-site-generator-like, data repository. If we go that route, it might be interesting to look at Gatsby which is really cool (and something I've wanted a reason to try out).

@khondula re: timeviz: Maybe, maybe not. When we have a single temporal coverage, I think it has low utility but it becomes really useful when someone includes multiple ranges (which we don't support just yet). Or if we're showing many datasets, each with their own temporal coverage, it'd be really nice.