I made a simple map of brazilian states in topoJSON and shapefile. Source for topoJSON is here in [Github](https://gist.github.com/ruliana/1ccaaab05ea113b0dff3b22be3b4d637#file-br-states-json) and visualization in [bl.ocks.org](http://bl.ocks.org/ruliana/1ccaaab05ea113b0dff3b22be3b4d637). Shapefiles can be found [here](https://github.com/ruliana/br-states).

![TopoJSON of Brazil’s states](/media/01M43VVN2672VWGJ30F8153TJH/hand-simplified-brazilian-states-1.png "TopoJSON of Brazil’s states")

*TopoJSON of Brazil’s states*

It’s pretty much the same as the project from [Carolina Bigonha](https://github.com/carolinabigonha/br-atlas), but with a minor twist: it was *simplified by hand* :)

If you take a closer look, it’s possible to see I simplified all the borders, removing details, but preserving the shape and precision as much as possible. The final file size is really *small*, no more than **8.5Kb** for the topoJSON file.

![Less precise, but smaller](/media/01M43VVNNSB44NX3BZ6YN6X0E8/hand-simplified-brazilian-states-2.png "Less precise, but smaller")

*Less precise, but smaller*

I made it for my [master’s course](http://up.mackenzie.br/stricto-sensu/engenharia-eletrica-e-computacao/areas-de-concentracao-e-linhas-de-pesquisa/) in [Mackenzie](http://portal.mackenzie.br/), we were doing data visualization and the maps I found had too much detail for simple choropleths or cartograms. QGis was used to edit it.

License is MIT, so use as you want.