Tutorials - Data Visualization Essentials

Networks

Networks are diagrams that are effective at showing intensity of relations among different nodes. Think of nodes as “links” which help establish the relation among properties. Unlike standard chart types, such as the bar chart, which is good for comparison, or line chart, which is good for showing progression over time, networks are effective at showing relationships. What differentiates networks from standard visualization types is that networks do not have a coordinate system. They do not need to be assigned X/Y position because what matters are how “links” are related rather than where they are positioned.

In this post we will go over the importance of using networks and create an easy-to-follow example using Quadrigram's drag and drop tool.

What are they?

Networks originally developed as part of mathematics. More specifically, networks come from graph theory. Graph theory studies “graphs”, mathematical structures that are used to model pairwise relations between objects (nodes). These pairwise relations can either be “directed” or “undirected”. Think of social media. In Facebook, for instance, whenever you add a friend you are establishing a relation between two people. This type of relation is undirected. Now think of twitter. Twitter is based of followers. So unlike adding a friend, following a friend indicates direction; someone is following you. This type of relation is directed.

Today, network graphs are used for just that. They are great for visualizing network information gathered in social networks such as Twitter and Facebook. Moreover, network graphs are also being used to visualize investor relations as well as many more data scenarios where there are related objects.

How do they work?

There are two ways to make networks. You can create networks either by using pairs table or by using an adjacency matrix.

Pairs Table

A pair table is a representation of nodes and relations. You need at least two columns, each row of the table represents a relation between two nodes.  

In the table above, there are three nodes (node1, node2, node3) and two relations (node1-2, and node1-3). Note that the pair table can also include another column with weights of relations among nodes.

Important: when working with pair tables it is required to label your first column as “Origin” and the second column as “Destination” for it to be recognized by Quadrigram editor.

Adjacency Matrix

An adjacency matrix is a square matrix containing all the nodes of a network, such as the following one:

You can plot a network using either pairs table or an adjacency matrix. This is the initial requirement for plotting a network. If you want to complement the network by adding visual attributes, such as color or size, then we need to build a second layout. This second layout is called “properties table”. With a properties table we can assign properties to nodes.

Step by Step Tutorial

Now that we know what networks are and how they work, let's see how we can create one using Quadrigram's drag and drop editor. You can follow along by watching the gifs posted below.

You can also download the data sets.

Note that there are two tables, “nodes” and “links”. The first one is a list with all the nodes, which can fir as the first column of our properties table. The “links” table includes a row for every relation. Therefore, we will be using a pair table for this exercise.

Nodes is your properties table and should look something like this:

Links is your pair table and should look something like this:

Using original tables, you can easily build a pair table and a properties table.

Properties Table in Excel

Pair Table in Excel

Now that we have the data let's load it in the Quadrigram editor and create a network.

Create: Quadrigram Network

Task 1: Make sure your data is correct

Everything looks good click OK.

Task 2: Choose Network and add your data

Task 3: Add Properties Table

We now have a network chart but there is a problem. The relations are not clear between the links. For this, we need to add visual attributes that define the relations between links. For this, we will use the properties table.

To add your properties table go to your Network Visualizer and select Style. In Fill Each Color Node select Colors by Category. Because we are looking to define the relation and not all the data, we are going to add the Type by selecting the Type Block and adding it to the Category Slot.

Almost Complete: Add Personal Touches!

Congrats! If you made it this far it means you have completed your network. It can be daunting at first but after a few attempts it will become natural and you will be on your way to become a data visualizer.

Now that we have the network chart, we can splash it with our personal touch by adding colors that fir our own personal style.

You can change the color of the relations as well as the color of each link.

P.S. If you would like to make networks and explore data visualization, we recommend Onodo. They have great data sets specifically for networks which can be downloaded and added to the Quadrigram editor. It is very simple: just change Source to Origin and Target to Destination and enjoy!