Network theory is a growing area of interest in modern science and has found applications in many areas, such as social interactions, economics, flavours in recipes and gene-disease relations. Many people might have heard about the small-world phenomenon, or the "six degrees of separation". In 1969, Travers and Milgram found that a message can be passed on to any other person by relaying it through the social network of people. As an experiment they asked participants to pass a message to a person they did not know, by using their friends and friends of their friends, basically a human chain. They found that the message only had to travel past a small number of people (6 on average) in order to reach its destination.

But what exactly is a network? A simple way of understanding a network is by assuming that a set of objects are connected by some sort of link. The set of objects may represent, for example, human beings, products, ingredients, diseases or brain regions, whereas the links are relationships or structural connections. In mathematical terms each object is called a**node** in a graph or a network, where the links are referred to as **edges**.

A graph or network can simply be represented by circles (nodes) and arrow (edges), where each edge can have a direction and weight. Directionality and weight can be understood in terms of friendship networks. Two people might see their relationship differently, in terms of how close they are or even if they are friends at all. Therefore a direction (who is friends with whom) and weight (strength of friendships) are useful information that can be incorporated in networks.

One of the simplest ways to characterise the nodes in a network is by their degree. The yellow node for example has 3 incoming edges and 4 outgoing edges, corresponding to an in- and out-degree of 3 and 4, respectively. Sometimes it can be useful to simplify a graph, for example by discarding the directionality information. In that case the yellow node would have a degree of 4. There are other measures that can be used to describe nodes or the complete network. The intuition behind some of the more popular ones will be introduced in the network measures section. Other aspects can be quite important as well, such as a network's topology or how we can create random networks. All these principles can help us understand what happens in networks that are too large to deal with by just looking at them, as they can be found, for example, in the human brain, which connects 100 billion nodes (neurons) through 100 trillion edges (synapses).

But what exactly is a network? A simple way of understanding a network is by assuming that a set of objects are connected by some sort of link. The set of objects may represent, for example, human beings, products, ingredients, diseases or brain regions, whereas the links are relationships or structural connections. In mathematical terms each object is called a

A graph or network can simply be represented by circles (nodes) and arrow (edges), where each edge can have a direction and weight. Directionality and weight can be understood in terms of friendship networks. Two people might see their relationship differently, in terms of how close they are or even if they are friends at all. Therefore a direction (who is friends with whom) and weight (strength of friendships) are useful information that can be incorporated in networks.

One of the simplest ways to characterise the nodes in a network is by their degree. The yellow node for example has 3 incoming edges and 4 outgoing edges, corresponding to an in- and out-degree of 3 and 4, respectively. Sometimes it can be useful to simplify a graph, for example by discarding the directionality information. In that case the yellow node would have a degree of 4. There are other measures that can be used to describe nodes or the complete network. The intuition behind some of the more popular ones will be introduced in the network measures section. Other aspects can be quite important as well, such as a network's topology or how we can create random networks. All these principles can help us understand what happens in networks that are too large to deal with by just looking at them, as they can be found, for example, in the human brain, which connects 100 billion nodes (neurons) through 100 trillion edges (synapses).

© 2016 M.D. Schirmer

www.markus-schirmer.com