Jurassic Journeys: Charting the web of Dinosaur connections through Topological Network Analysis in Power BI

What is Network Analysis?

Network Analysis is the study of relationships between a group of entities, let's say social relationships, network of stars/ planets/ constellations, transportation network, network of routes, and so on.

Network Analysis is conducted to understand functioning of the one system as a whole. It helps to identify patterns of a specific group (considering a group as one entity) and studying the behavior of those entities with respect to other entities in the same network. It answers questions such as 'What will be the impact on a team if one employee leaves the organization?', 'How will news of an advertisement spread out in a community?', 'How will the network behave if one or few nodes are added or removed?' 'How will the network adapt itself if changes happen?'

Network analysis can be carried out when there is a source node and target node and they are related to one another in some way. Network can vary from a very weak relation to a very strong relation.

There are many types of Network Analysis and each one has its own goal to study a group or entity as a whole.

In this study, I have taken historical dinosaur data which looks like this:

Now to create Edges and Nodes, we have to split this table in two separate tables. The first table has Dinosaur species and related information and second table has information related to places where they lived in. Both tables are connected with a common column Dinosaur Name.

After plotting Continent and Dinosaur type respectively in Source Node and Target Node in Network Navigator visual, we get the graph as:

We can observe with this diagram how different types of Dinosaurs are connected with the continents.

However, if I remove Asia from the network, we can observe that network is smaller now.

Similarly, if we plot Countries and Dinosaur Names in Network Navigator, we get the following visual. As Dinosaur names are unique in the given dataset, only the dinosaurs which are living in a particular country are seen connected to that country.

The above is the overview of the complete picture and is difficult to read in first place. After zooming in a specific country we are interested in, we can better study about the central point and the entities holding that point.

Now, lets observe how different dinosaur species are connected to different countries. The overall picture is :

This shows an interesting network how same species are present in different countries, whereas some of those are not related at all.

Let's take a closer view of one network.

There are three species of dinosaurs which are common in between Mongolia and China and strong relationship is found. However, if I remove the species between China and Mongolia, the network is changed to this:

This analysis is called as 'Node Removal Analysis', where one or more nodes are removed and the results are observed. Now, the intermediary node is USA, which is altering the shortest path between Mongolia and China.

By this analysis, we can clearly see how the changes in local nodes can have global implications.

To conclude, by removal or addition of network nodes, networks can adapt themselves. We can study dynamic nature of network structures which can answer questions related to a group.

Thanks for reading!