Let’s talk about Relationship Cartography in general. What is it? Well, I’m not 100% sure yet as I just made it up. The thing is, I’ve been looking at so-called Social Graphs, Commercial Graphs, Economic Graphs and more lately. And I’m struggling to find a unifying theme for these various sub-types of relationship visualizations. Maybe I’ll go into these various types in detail later on in a follow on post Types of Relationship Graphs, but for now I’d like to just blather a bit about graphs in general. You’ve probably heard of them. And maybe done a little research. Still, it’s sometimes useful to take a high level conceptual view of things to understand the parts a little better.

So What’s the Deal with these Graphs?

Well, they’re of course not graphs as we typically remember them from math class. I’m not even sure exactly how it became a meme. Maybe because ‘graph’ was the closest thing one could get without saying it’s a Data Structure of Connections. (Which would of course just result in the acronym DSOC and be yet another standard somewhere.)

So if they’re not cartesian graphs where we can plot stuff on X, Y, Z axis, just what are they?

These graphs are essentially relational representations. These representations can manifest themselves in many ways. You’ve typically seen them drawn as maps of a sort; often cluttered with relationship lines. If you have the opportunity to use various visualization tools, you should be able to zoom in and out on sections of interest. But these graphs may also be represented by pure numbers. (Showing things like ‘nodes’ which are the objects themselves, and relationships, which express how nodes relate. Many additional data points apply such as strength of relationships, which can in turn be inferred by a variety of means from frequency of contact to direction of contact to, etc. etc.) The study of these types of graphs is referred to as graph theory.

graph theory is the study of graphs, which are mathematical structures used to model pairwise relations between objects. A graph in this context is made up of vertices, nodes, or points which are connected by edges, arcs, or lines. A graph may be undirected, meaning that there is no distinction between the two vertices associated with each edge, or its edges may be directed from one vertex to another  — Graph Theory – Wikipedia

You can find out even more about the relationships within these graphs in another Wikipedia Article Graph (discrete mathematics.)

Do Today’s Labels for Graphs Make Sense?

These graph types still seem a bit ill-defined. And probably mis-used in a lot of cases. For example, is Facebook really your Social Graph? Or is it a somewhat accidental collection of people you happen to know due to where you grew up, went to school or whatever? Does it really include all of those you consider friends or acquaintances with whom you interact with in the real world? For example, does it include the parents of the kids you carpool with? They may not be friends per se, but they’re certainly a part of your social graph that you interact with more than most of your declared friends. What about your buddies on your hockey team? Maybe some of them are Facebook friends, but all of them? You can come up with many other examples. But what it really comes back to is that something like Facebook is really your Friends & Family Graph, but hardly begins to cover what is likely to be a much larger range of social interactions. It still may have tons of value, but it’s not really the whole thing. Maybe this doesn’t matter much, but perhaps the naming of things and their definitions do matter.

The fuzziness of new definitions like this is to be expected. In anything new, we struggle to define things and sometimes our words or phrases come up lacking. In other cases, we end up just changing the meaning of the words or crowding out old meanings. (Type “java” in Google and see how far you have to go down to get to coffee. When you search for “cancer” do you mean astronomy, astrology or oncology? For “bass” do you mean the fish, the shoe company or a bass guitar? It seems disambiguation has long fallen out of favor with top search engines so we’ll likely see more linguistic prioritization brought on by machines. But that’s a whole other topic.)

I’d love to have used the phrase Social Cartography instead of Relationship Cartography. However, that actually has a special meaning for the Sociology Crowd. Which is fine. They’ve actually defined it somewhat clearly.

Feeding the Graph

Your digital exhaust from all manner of behavior continually dumps into the Great Maw of the graph space. Who gets it and for what use is the only question. As most people know by now, marketing companies can pretty much track you wherever you are on the web. Though they may lack the ability to see into particular services for some of your relationships, they may be able to infer a lot more than you think. Especially if tied to offline behaviors from your physical frequent shopping card to credit card transactions.

The idea is that the data gathering process may be participatory, or not. That is, you may explicitly provide information to others or just have your data points ‘sensed’ by some means. Mostly of course the gathering process is not participatory. The closest you get to explicitly offering to provide info is agreeing to the 100 page Terms of Service for whatever online products you’re using.

From an academic point of view, up until the rise of large scale social networks, sociologists were sorely lacking in data to test a wide variety of theories.  The primary barrier to data gathering was and remains the cost to do so in the real world. There’s also the issue of affecting your own experiment with what for most experiments would be self-selected samples as participants. Now it may be true that social networks hardly represent the entire universe of interactions among people. Nevertheless, they’re a treasure trove of data for those social scientists allowed a peek behind the curtain. (Of course, one could leave academia and just go work for a social network as a sociologist / data scientist. Not too many years ago it would be practically beyond imagination that such a role could be a high paying tech job.)

So What’s the Point?

The point is that labels matter and you should be careful with them. I probably have no hope of changing the abstract idea of Social Graphs to Relationship Cartography, but perhaps if you’re working with this technology or considering marketing programs, you try to think of things this way.

Relationships are more nuanced than what is likely wholly represented in particularly expressed explicit system graphs. (Though they may be able to infer implicit relationships.) OK, that bears some explanation. You know how Facebook or LinkedIn sometimes suggests people you may know? They obviously do that based on some kind of signals. How many friends of friends do you have in common, cluster analysis and so forth. So your Friends on Facebook are those you have explicitly defined. And social scientists, programmers, marketers may infer other implicit relationships based on your various graphs or other behaviors.

As a designer, programmer or marketer, try to be clear about what problem it is your trying to solve and how you use relationship data to solve it. For example, in advertising context clearly matters a lot. Hence the success of pay-per-click search and native advertising. If you get the context wrong, you fail. If you get too creepy with use of context, you also fail. Whenever your using Relationship Cartography, Social Graphs, whatever you want to call it, use it to understand your customers as best you can before going right to messaging or other goals. Doing so will help inform your choices.

Reading for Extra Credit

Social Graph Problem
The open graph
Social Network Analysis
Owning Your Social Graph
The graph API, nodes, edges and fields
The Social Graph is Neither
Explaining what the “Social Graph” is to your Executives
Sociometry
Facebook’s “Open Graph”
Social Graphs for Business
Business Social Graph
The NSA Knows
Social Media Research Foundation
TouchGraph Navigator (for visualizing relationships)