Graph visualization/analysis tool?

Question

Since the Harvard Business Review claimed that Data Scientist is The Sexiest Job of the 21st Century, graph analysis and big data are becoming very popular topics at my university.

I'm familiar with graphs algorithms, but we usually program our own task-specific tools. Are there any good applications to visualize and analyze graphs?

Answer 1

There are actually quite a few applications for visualizing and analyzing graphs:

• Gephi and Cytoscape are two well-known open source applications that support large and complex graphs.

• If you're mainly interested in visualizing graphs, have a look at Graphviz, which is an absolute classic.

• You can also use R or commercial tools like Mathematica if you're more interested in the statistical and analytical aspects (see also this question over on Stats SE).

Answer 2

One that just appeared is https://plot.ly. There are many more. Which program is most useful to you depends on a lot of factors. If you are technical proficient, you might like Weka (http://www.cs.waikato.ac.nz/ml/weka/).

Answer 3

There are also a number of network analysis packages for the open source R language, such as network and sna, and igraph, all of which have some viz capabilities. R can also be a good environment for general data manipulation tasks.

Answer 4

d3.js is also a good javascript library manipulating data. Though most commonly used for building visualizations, you can do any sort of data-driven manipulation in the browser.

vega is slightly higher-level visualization grammar built on top of d3.

Answer 5

for python i recommend igraph and networkX, for parallel computing graphX.

Answer 6

If your main focus is to visualize the data and have graph metrics I would recommend this list.

If you want an off the shelf package

• Gephi - Desktop application, Open Source license
• Cytoscape - Desktop Application - Open Source license
• Pajek - Desktop Application, Free for non Commercial use (old but still good - I've seen papers using it just few days ago)

If you have any programming skills

• KeyLines - Javascript Toolkit, SNA metrics, Commercial license
• VivaGraphJS - Javascript Toolkit, few SNA metrics, Open Source licence
  • As mentioned in the page linked, the library will close in faviour of ngraph in the future
• igraph - R/Python Toolkit, more focused on SNA rather than visualization, Open Source license

Others or "Why you didn't mentioned this/that?"

This is a list of more libraries that will let you do only a part of what you're asking or there are strong constraints on the use:

• NetworkX - Wonderful for SNA, Open Source license
  • it doesn't visualize by itself, so use it in combination with SigmaJS or some other Python library
• Linkurious - Easy to use, with SNA, Commercial license
  • you need a Neo4J backend in order to use it
• d3.js - Wonderful to visualize data, but it has not built-in SNA metrics

Disclaimer: I'm in the KeyLines team.

Answer 7

There are a few solutions to visualize graphs : D3.js, Sigma.js, KeyLines, Gephi, Linkurious, Neoclipse, Neovigator.

Here is a table that compare some of these options : http://linkurio.us/comparative-study/ And a quick presentation about the different approaches to visualizing graphs : http://www.slideshare.net/Linkurious/graph-visualization-options-and-latest-developments

Linkurious is specifically designed with ease of use in mind. You can use it search, explore and visualize graph data easily.

Disclaimer : I'm a co-founder of Linkurious

Answer 8

My team builds https://vida.io a tool for creating data visualization templates. We support d3.js templates. You can see a lot of examples on our site:

https://vida.io/explore

For simple chart/visualization, you can use Google visualization tool:

https://vida.io/pages/google-charts

Answer 9

Quadrigram (www.quadrigram.com) has good graph visualizations in both 2D and 3D. They are relatively easy to set and publish.

You can also combine them with Maps and other visualizations Checkout this example here.

Answer 10

There are a number of packages in the R language very useful to data analysis/visualization. Hadley Wickham has developed lot of interesting tools to make these task easier. The recent bigvis package is very promising.

Answer 11

Check out Tableau / Tableau Public. It was used to build this : http://public.tableausoftware.com/shared/QZPQ8DCR9?:display_count=no

It has a loot of tools for building a visualization that lets you explore / drill into the data. It can let you build a sort of interactive experience that brings a lot more meaning to your graphs and charts.

The public version is free, student / non-profit pricing is fairly reasonable if I recall.

Answer 12

The answer has changed a bit nowadays:

• Analytics: For accelerated computing, cuGraph (GPU-accelerated Python) as part of the broader RAPIDS Python GPU ecosystem. For many diverse algorithms, Neo4j has great extensions, and for potentially more scalable experiences, TigerGraph and GraphX. Graph neural nets are emerging as an exciting area but no clear best practice yet.

• Visualization: Graphistry (bias as the founder ;-)) is the only end-to-end GPU-accelerated visual graph analysis tool and runs in browsers even 5+ years old. Likewise, for use by analytics teams, it is also gaining popularity due to Python notebook support and built-in visual analytics (1K+ GitHub stars). D3/Linkurious/KeyLines/etc. are good for small graphs, with D3 being free. However, for bigger ones, all of these are still much slower than Graphistry from 5 years ago, meaning you or your user's browser will crash on basic tasks like event data, and have little for already built-in visual analytics similar to desktop Gephi or beyond