Google Maps API Key Warning - D3.js alternative?

[photomedia]

GeoChart throws a warning about requirement for a Maps API key:

This is a relatively new and unpopular policy change at Google with respect to the Maps API.

The Map visualization (for Origin of Downloads report) still shows up, for the time being, but this warning means that Google will probably keep up the pressure for us to get API keys (which now requires a method of payment/CC/account on their cloud platform).

I think we should seriously consider switching over to D3.js (https://d3js.org/) implementation, at least for the map visualization. I don't think it is feasible or sustainable for an open source repository platform like EPrints to be requiring to add a credit card paying account on Google to visualize some usage stats.

I suggest D3.js because it is popular and BSD licensed, but perhaps someone knows if there are other better alternatives?

[Monica-Wood]

@photomedia I would be interested in moving away from Google implementations. Let me know if you would like to pursue this further.

[photomedia]

@RedHaxor yes, I'm interested to pursue this further. I think that the first step would be to choose the open source library that we want to use instead of Google for the graph visualizations. The bar charts should be relatively straightforward, a number of options exists for that (D3.js is one option), but the map visualization of countries is the one that is more difficult to replace, but I think that D3.js can be used for that too.

[Monica-Wood]

@photomedia, I've only had a very quick look at D3.js, but it looks like it could work. Best thing to do is a small test to see if each graph type will work with the eprints data and within the existing template.

[photomedia]

@RedHaxor A complementary alternative (this one uses D3.js as a dependency) is to use the Vega Visualization Grammar https://vega.github.io/vega/ or Vega-Lite https://vega.github.io/vega-lite/examples/ I don't have much experience with this one, but the example gallery looks to include any visualization that we need: https://vega.github.io/vega/examples/ The deployment looks simple enough, and the community around this one is large. Vega is also used inside of WikiPedia.