There is a very special case of fixed layout that we need to talk about. This special case is when the fixed layout is actually a map or spatial locations. So, in this case, the nodes are positioned in a way that their position represents some spatial location. As you can imagine, this is a very important method and a very important situation because there are lots of datasets out there that actually have this structure. There is a geographical representation. The nodes are placed in positions that represents some geographical coordinates or spatial coordinates, and there is something that connects these nodes. Very typical example is migration maps for instance. So, datasets that describe people migrating from one region to another of the world, or a specific state, or country. Well, that's a classic. Here is an example. So, this map has been designed by Ilya Buoyandin who is actually one of my former students. It depicts exactly what I just said. It's a migration map within Switzerland. As you can see, the nodes are placed exactly in some specific regions. They actually do represent some regions, and you have people flowing from one region to another. The flowing is represented by the edges. The edge thickness represents the amount of people that are flowing from one region to another. In this case, you can also see that there is a directionality. The way directionality is been depicted in this map is using these half arrows. So that for every pair of nodes, you actually have two edges, one that goes in one direction and the other that goes in the other one. So, one very common problem with these kinds of maps as well as in general with node-link diagrams, is that you can create a lot of clutter. As soon as you go pass beyond a few hundreds of nodes and edges, typically you have a lot of clutter. So, this is an example, right? So, that's another origin destination map as people used to call it or sometimes they call it flow maps, and it represent flights within Europe. Frankly, it's a mess. You can't really distinguish much in this map. So, what can you do here? Well, interestingly, one thing that you can do is to use exactly the techniques that I described before, which is edge bundling. So, here on the right, you see exactly the same dataset represented with the same technique where the only difference is that edge bundling has been applied. So, this is very, very powerful. Here is another similar image. This is been created with a library that is called QGIS. What is really interesting here is that in addition of using edge bundling, color is also used to represent directionality, right? So, you see that there are color shades and it represents what is the origin and what is the destination. So, this is a general problem that I've introduced before. Every time you have directed graphs, you have to decide what is the best way to show directionality. As we have seen in these last two examples, one way of representing the directionality is by using arrows, and other way is by using colors. There are many other possibilities here. These are some of the most used ones, and they tend to be very effective.
