Best Network Visualization Software Tools for Data Analysis

Posted Monday, June 6, 2022

Network visualization software can be used for extracting relations and patterns from data. Some of the insights can also be used for feature extraction to improve machine learning models. Below we present a comparison of the best tools and software libraries available today both for beginners and tech-savvy users.

Most of the network visualization and analysis tools available today fall into one of the following categories:

1. Online applications and analytics platforms (e.g. InfraNodus, Graph Commons, Rhumbl)
2. Desktop applications (e.g. Gephi, Pajek, NodeXL)
3. Python and R-based software libraries and packages (e.g. NetworkX, iGraph)
4. Javascript libraries (e.g. Cytoscape, Graphology, Sigma.Js)
5. Graph databases (e.g. Neo4J, TigerGraph)
6. Excel plugins (e.g. NodeXL)

If you know how to program and need a high degree of customization, you can go for for Python or R-based software packages. These can be used to build your own applications and also provide maximum control over your data. For more advanced users, installing a graph database, such as Neo4J, also gives you the access to the analytical tools that are part of the package.

If you are a data scientist, you might want to use a desktop application or an online platform. These can save the development time and provide direct access to graph analytics and various network metrics.

Online Network Visualization Tools: InfraNodus, Graph Commons, Rhumbl

If you would like to build a graph using your browser and easily share it with the world, you can use InfraNodus, Rhumbl, or Graph Commons. Rhumbl and GraphCommons are a bit similar in a way that they have an interface that enables you to customize the look of your graph. However, they provide very rudimentary analytics features. This is where InfraNodus truly stands out, as it provides not only a graph constructor, but also advanced network analytics that are lacking in the other tools.

All of these tools either offer free accounts or trials, so you can try them online. Paid accounts are relatively accessible, starting at €9 a month, and some of these tools also offer an API. All the tools let you share the data using embeddable iFrames, URLs, or plain images. Only InfraNodus offers high-resolution vector-based export, which is suitable for print publications.

In terms of the usability, it really depends on your preferences, we describe the peculiarities of every tool below:

InfraNodus: A Universal Tool for Network Analysis and Visualization

InfraNodus is a network visualization and analysis tool. You can import your data from a spreadsheet, edit the graph manually, or use its live text-to-graph editor to quickly create your graph. InfraNodus will automatically provide graph analytics, such as community detection and node influence measures normally available in advanced tools such as Gephi or NetworkX. However, due to the fact that the look of the graph is defined by the analytics, you don't have such a high degree of customization as you'd have in Rhumbl or Graph Commons. On the positive side, InfraNodus offers AI-based recommendation system that traverses the graph and recommends new connections that bridge the structural gaps in your data. Every piece of data you obtain using InfraNodus can be exported as a spreadsheet or JSON for further analysis. The graphs can also be exportd in a high-resolution vector format (SVG), which is suitable for large-format print.

Graph Commons: Good for Visualizing Small Graphs, Bad for Analytics

With Graph Commons, it can take a while to build a graph, because you have to specify a name and category for every node that you want to add. This provides a high degree of customization, but may also get slow for the big graphs. Graph Commons can be useful for building small graphs that illustrate a certain aspect of a phenomena, not more than 30-50 nodes. You can also import the data, but you have to save that data in a very specific format as a spreadsheet, compatible with Graph Commons template, so it may be difficult to convert your existing data into their format. Graph Commons also has a built-in presentation tool, so if you want to make a slide-show using their graphs, it might be a good option. Graph Commons do not offer graph analytics, so you don't really harness the network science insights from their tool, but you do get a nice, easy-to-read visualization.

Rhumbl: Highly Customizeable, Suitable for Big Graphs, but Lacks Analytics

Rhumbl is a tool that can be used to build both small and large graphs. However, its interface is pretty technical, so it will probably be accessible only to the people who understand how network graphs work. There's a huge amount of detail: you can describe each node and relation, but that might also get overwhelming. You can choose the template for your graphs, and you have high degree of control over the colors and design. However, they don't offer any graph analytics, so just like it is the case with Graph Commons — you get a nice visual representation, but you don't harness the power of network science to its fullest extent.

Desktop Applications: Gephi, NodeXL, and Pajek

If you don't mind installing a desktop tool and spending some time to learn how it works, you might be intersted in Gephi, NodeXL, or Pajek. These are advanced graph analysis platforms that provide most of the metrics normally used in network science. Here are some interesting facts about each of them:

Gephi: Great for Analysis and High-End Visualization

Gephi is a wonderful tool that has both network metrics and advanced visualization customization. You can calculate nearly every measure that you can think of and align the nodes in many different ways. You can also design the look for your graph once it's exported as an image: e.g. choosing the shape of the edges and of the nodes, defining the colors for each community, etc. Gephi is free and is cross-platform as it runs of Java. You can even run it in server mode, so it can be integrated into your existing applications using its API.

NodeXL: Best Suited for Social Network Analysis

NodeXL will only run on Windows and its interface is a bit old-school. However, it can still be a good choice for performing social network analysis. NodeXL seems to be not so frequently updated these days and they offer a confusing range of options: for example, their browser-based version is nowhere to be found, while the Pro desktop version is a bit cumbersome and slow.

Pajek: The Old Classics

Pajek is perhaps one of the first popular network analysis tools available on the market. It only runs on Windows and was last updated in 2009, so we keep it here for historical purposes only.

Software Packages: NetworkX, iGraph, Graphology

If you are a tech-savvy who likes programming, you might be interested in Python- or R-based NetworkX and iGraph or Javascript-based Cytoscape and Graphology. You would obviously need to program your use case manually, but all those libraries have multiple examples and templates. All of these packages are free and open-source.

NetworkX: The Popular Python Classics

NetworkX is the most popular Python network analysis package with some visualization capabilities. It is frequently used in various types of applications and is the industry standard. It is free and open-source, you also benefit from active developer community and multiple examples that can help you build towards your use case.

iGraph: Faster Processing

iGraph is perhaps better suited for building large graphs as its base code is implemented in C, so it may be faster than NetworkX if you are concerned about the speed or have to work with the large graphs. Here is an interesting benchmark comparison of the both tools where iGraph is definitely a winner. Some people complain that iGraph is harder to use and has less documentation than NetworkX though.

Cytoscape: Javascript Network Analysis and Visualization Powerhouse

Cytoscape is a javascript-based library that was initially built for analyzing biological networks. So it can handle vast amounts of data, has multiple network metrics, and boasts its own visualization package. Some of its metrics is used for graph analysis in InfraNodus.

Graphology: A Light Graph Analysis Library

The advantage of Typescript-based Graphology is that it's the newest package from the list, so it's really lightweight and less bloated than the others. It also has a tight integration with the amazing Sigma.Js network visualization package, making those two a true gem when it comes to visualizing and analyzing both small and large networks.

Graph Databases: Neo4J, Titan, OrientDB, Tigergraph

If you are deeply immersed into network analysis, you might want to install your own nosql graph database. All of the graph databases available today offer their own graph analysis algorithms, which can be ran on vast amounts of data. You would need understand their query language and know how to program, but for large-scale applications these will provide the highest level of customization.

Neo4J: Widely Used and Reliable

Neo4J is one of the oldest and most popular graph databases. It offers advanced graph analytics and easy-to-use Cypher graph query language. They have a community version that is free and also offer their own hosted instances, which is great if you want to set up a test DB to play with. The default visual interface is a bit laggy, but if you're happy to code, you won't be disappointed.

TigerGraph: Built-In Machine-Learning

TigerGraph is the most hyped graph database today as they offer built-in machine learning capabilities. They also claim to be the fastest graph database on the market.

Network Analysis Tools: Similar Functionalities Graph

Here is a graph of the most interesting tools for network visualization and analysis represented as a graph. Choose the functionality you need and you will see the tools that offer this functionality as well as the relations between them. It is made using the InfraNodus network mind map feature:

Try It Yourself

You can perform network analysis and visualization on your own data using InfraNodus:

Log In    Sign Up
 

 

Custom Service: Network Visualization

Please, let us know if you have an interesting proposition, use case, research project, or service request.

Contact Us