Introduction

The bottlenose dolphin is a very intelligent social creature. Just like humans, dolphins group have their own social connection with each member. They communicate with their group by ultrasonic, which can help them exchange information and divide their work and make decisions.

This visualization work shows a sample of a social network of 62 bottlenose dolphins. My data is downloaded from a database website “Kaggle“. This data’s creator is Maria Ilina. 

Inspiration

The inspiration I found to prepare this lab is a network visualization which shows connections of people’s philosophic idea in history. The author makes the nodes bigger if this person’s idea influenced more people. As we can see clearly, the biggest nodes are very famous people, such as Plato and Aristotle, which means they made a very great contribution to human progress. The color of nodes also represents different genres of that author: Green node means antiquity & enlightenment philosophers. Red node means 19th/20th-century philosophers. Pink nodes are enlightenment, Yellow is 19th/20th-century authors, and Orange is fiction author, Purple stands comedians.

Although the color assortment is a little bit confused to the audience, the information this graphic wants to show is very clear and understandable because the contrast of the node’s size is very impressive to the audience. That’s why I think my own visualization can follow this rule to display the data.

Method

The network data visualization is very complicated and specific. They have Node, Edge, Direction, and some other attributes which are totally different from the data type I found before. This type of visualization is very efficient to show how things are interconnected through the use of nodes/vertices and link lines to represent their connections and help illuminate the type of relationships between a group of entities.

The tool to realize my network visualization is Gephi, which is especially for data analysts and scientists to generate the data network graphs. There are many features in this software like representation, manipulate the structures, shapes, and colors to reveal hidden patterns.

In order to shows the most connective dolphins in this group obviously, I plan to display the nodes in different size and ranked their size by the connection amount of them. By using the ranking feature in Gephi, I can generate the different size of nodes based on its connection number. The rank of color also reflects the different densities of the connections. (I used blue as the theme color because I think it is related to “feelings of dolphin”.) The node is bluer if its connections are more so that we can find out more connective nodes easily.

In order to show the structure of the dolphins social network, I use “Yifan Hu” layout to visualize the network. In the image above, we can see this network is separated into overall two clusters. The right one is bigger and more connective, the left one is smaller. Through observing this network, we can find that this social network consists of two subsets and a bridge between these two subgroups. If we want to know what dolphins are acting the bridge role in this system, this version may be more convenient for us to find out.

Reflection

Because my data is relatively small and the attribute of my data isn’t complicated, the display of this data is easy to understand. The feature I used in Gephi is mainly about the ranking display and the previews setting. I still feel Gephi has limitations because the final visualization is not interactive. I expected that if I hover the mouse to a node, the image will highlight its connected nodes.

I read an article said that the dolphin’s social network is mainly based on age, sex, reproductive condition, family relationships, and association histories. So I think if I can know, such as the sex or the age of the dolphins in this group, I can use different colors to show their sex and use different shape to show their ages. If so, the data will be more interesting and informative.

Before using this data, I tried to find some interesting datasets to take better advantages to Gephi to visualize the connection. However, some network datasets are too big or small to visualize. My computer cannot process the network that contains over 2000 nodes. Also, some datasets only have the id number of each node, which is hard for me to understand. So for this work, I really spent lots of time finding proper data. I am glad that this dolphin one is relatively clear even it is simple and small.

Reference

D. Lusseau, The emergent properties of a dolphin social network, Proc. R. Soc. London B (suppl.) 270, S186-S188 (2003).

D. Lusseau, Evidence for social role in a dolphin social network, Preprint q-bio/0607048 (http://arxiv.org/abs/q-bio.PE/0607048)