Getting started with maps in Python | Hacker News Reader

Getting started with maps in Python

maxberggren.github.io

121 pointsmaxberggren15 comments

Nice walkthrough! However, I'm surprised they're using basemap for this. Basemap is still maintained, but it's more or less been superseded by Cartopy. (Also, Cartopy should support the Winkle Tripel projection, as Proj4 supports it.)

Edit: Scratch the Winkel Tripel projection support. Proj4 doesn't define an inverse method for that projection (it's non-trivial), so it can't be used correctly by cartopy.

I've been working in this area and I do generally prefer Cartopy over Basemap. Our team recently made this switch. Cartopy is more Pythonic and has a better software design. It has nice abstraction for dealing with and using different map projections. It also handles seam issues well (I cannot recall if this is a problem in Basemap or not). All that said, Basemap has features that are lacking in Cartopy (e.g., can you label lat/lon lines on non-XY projections in Cartopy? The last time I looked, this was not possible.) Also I agree that Cartopy documentation is somewhat lacking in some areas. But going forward I would probably choose Cartopy, notwithstanding Jeff's great work over the years.

I'm the author of http://sensitivecities.com/so-youd-like-to-make-a-map-using-... and make maps using Python all the time.

I'd love to switch to Cartopy, but its documentation isn't comprehensive or well-presented. I check back every six months or so, and then forget about it again.

This one really helped me back when I needed choropleth maps. So thanks!

I myself prefer Kartograph, which exports SVGs and I think yields very pretty maps: http://kartograph.org/

Kartograph is quite nice. I need to play around with it more.

However, it's aimed at a very different audience (web-based interactive rendering of vector graphics, as opposed to desktop-based interactive rendering of scientific data). For example, as far as I know (I could be very wrong about this), there's no way to display image/raster data in kartograph.

That having been said, kartograph is a very slick library, and its built-in clustering methods for point data are particularly nice.

OP here. Cartopy is actually news to me! Thanks for the info! HN at its finest.

#Edit - sorry didn't see the link in the original article. I was asking about notebooks, and you linked one already.

I use https://github.com/andrea-cuttone/geoplotlib in my work and highly recommend checking it out.

The lib uses openstreetmaps as background, can display scatters, heatmaps, shapefiles, calculate and display voronoi tesselation, and does a ton of other things.

nice - here is a cartopy version .. https://gist.github.com/darkblue-b/6d9aab096ffe9bce3a52

(that dataset and more examples are freely available from OSGeo) http://live.osgeo.org

Does the author even know Python? Because in Python maps are called dicts!

FYI the submitted article is about cartographic maps.

I think (hope) it was a joke.

at least someone gets me!

lol, yes.

Reply on news.ycombinator.com