Weighted Voronoi? Different distance algorithms?

[mbostock]

Related d3/d3#2401. For example: Manhattan, Chebychev, and Minovsky.

[BBischof]

I'd also like to echo the request for support for Weighted(additive and multiplicative) Vornoi.

[Fil]

Seems that the reference algorithm can handle weights
https://en.wikipedia.org/wiki/Fortune%27s_algorithm#Weighted_sites_and_disks

[jarmstrong2]

Would also like weighted support if possible! Thanks!

[Fil]

If we could include geoDistance() in the pack then it would help solve Fil/d3-geo-voronoi#1 too!

[nitaku]

@mbostock Any pointers on which parts of the code are involved? I'm also interested in having custom distance functions.

[mbostock]

@nitaku No pointers, sorry. If I knew how to do this I would have done it already.

[jarmstrong2]

Mike, what is the original algorithm for d3's voronoi based on? If it is in fact Fortune's algo, there is a weighted variant available.

[mbostock]

Yes, as @Fil mentioned. This library is an adaptation of gorhill/Javascript-Voronoi, which is an implementation of Fortune’s algorithm.

[aubergene]

I'd love to see these additions. A while ago @jasondavies made an example weighted voronoi power diagram, he's since replaced it with an image. I think it was written with an older version of d3 which I didn't use Fortune's algorithm

[nitaku]

Maybe @gorhill could help us to understand which parts of the implemented algorithm are involved. I had a look at his github, but I was unable to identify the code about distance metrics.

[Kcnarf]

I've found an interesting bl.ock implementing a Weighted Voronoï algorithm in order to produce a Voronoï Treemap : https://bl.ocks.org/mkedwards/759e719eefe36cf9c8ab.

I guess the algorithm uses a 'power distance' as the tessellation uses straight lines. Also, I've not yet been able to identify the underlying algorithm, nor if it is a port from another language.

Nevertheless, it could be a good starting point.