Dynamic initial frame

[samwhite]

badger-map/index.html

Line 177 in cc5e81f

map = new google.maps.Map(main, {

It could be cool to take the data from the spreadsheet, figure out the outermost locations, then open the map with the zoom/centre such that all pins are visible

[charypar]

It would, but maybe it needs to be a bit probabilistic, otherwise trolls like Chris Baron, who place themselves into Zanzibar will make the initial frame unusable.

Maybe what we could do is do an estimate a 2D Gaussian over all the points and then cover 2 or 3 sigma of the space...? That can in fact be done in the spreadsheet probably 🤔

[charypar]

Upon some experiments, using percentiles will do the trick. But even simpler, given the zoom levels are discreet, it might be better to estimate the centre and guess a more appropriate zoom level...

[charypar]

The current centre point is Charing Cross - 51.507, -0.128. The median of all the data points at the moment is 51.527, -0.081.

Only very slightly different, i.e. good estimate, so that may be worth doing to make the map a bit more universal.

[charypar]

Bounds fitting can be done with https://developers.google.com/maps/documentation/javascript/reference/map#Map.fitBounds

A good bounds approximation seems to be from somewhere around 1.8 percentile to about 98.2 percentile in both coordinates. Due to the zoom level fun, I imagine anything between 1 and 2, and 98 and 99 will probably do ok.

This looks like a sensible way of calculating the percentiles: https://groups.google.com/d/msg/microsoft.public.excel/DPVw83CE0r8/4hfzQKbuav8J

[BadgerBaron]

You can't pinpoint Randy Mclovin - he's everywhere

[robbiemccorkell]

Could just delete the shit data

[charypar]

Possibly made obsolete by #17, which zooms on current location, if geolocation is allowed.