Network overview, discovery and exploration for Microsoft Excel.
Network graphs provide a great way to explore and visualize relationships between social media entities. Each actor is represented by nodes, and their relationships are depicted by line connections. One classic tool that allows you to explore social relationships via network graphs is NodeXL. And, following the launch of its advanced version last week, now you delve even deeper into your data.
NodeXL basic is a free, open-source template for Microsoft Excel that allows you to create a network graph within the familiar context of a worksheet.
The NodeXL template supports import data from a variety of sources including Excel workbooks, UCINET files, Pajek files, GraphML files, Wikipedia, Facebook, Twitter, and other social media sites. Additionally, if you wish to import data from an unsupported source, NodeXL contains a plug-in architecture that can be used to do so.
Once data has been imported, NodeXL harnesses a number of community detection algorithms to easily identify connections between social media entities. These can then be explored through common graph metrics, including degree, centrality, and clustering coefficient.
To make the data meaningful, NodeXL contains several visualization options in order to highlight certain relationships. The default graph utilizes a Fruchterman-Reingold algorithm, but the template also offers other layout options, such as Circle, Spiral, Grid, and Sugiyama graphs, and more advanced users can manipulate the vertices in a manual layout.
As an open-source tool, NodeXL users have the added benefit of leveraging off community resources - most notably through its Graph Gallery. Here, users can upload and download NodeXL files. A particular graph's parameters are available for download in a NodeXL options file, which mimics the colors, vertex shapes, edge widths, layout algorithm and other characteristics of the chosen graph for your unique dataset. Some users also make their datasets available for download.
On 12 October 2015, NodeXL released an advanced version of its template called NodeXL Pro. For a fee, users can utilize additional analysis and import features, such as access to the full Twitter API, data from Facebook groups and fan pages, the capacity to import and export to GraphML, advanced centrality analysis and automation features. Further capabilities are already being planned for the Pro version, including a sentiment analysis tool due for release in November 2015.
An example: MH17
Following the reconstruction of MH17 last week, Marc Smith conducted a NodeXL analysis of MH17 mentions on Twitter. His graph utilizes a Clauset-Newman-Moore cluster algorithm for the vertices in conjunction with a Harel-Koren Fast Multiscale layout. Within this, Smith used an edge for each "replies-to" relationship in a tweet, an edge for each "mentions" relationship in a tweet, and a self-loop edge for each tweet that is not a "replies-to" or "mentions". From this analysis, it is possible to deduce the top influencers on Twitter in the MH17 discussion, and other conversation dynamics. For journalists, these kinds of analyses are particularly useful to identify sources, explore alternative story angles, or report on the social impact of an event.
Image: An overview of Marc Smith's MH17 NodeXL network analysis
Image: Top influencers in Marc Smith's MH17 NodeXL network analysis
View the interactive version of the MH17 network graph here.
Want to learn more? Check out the following tutorials:
- NodeXL's comprehensive list of social network analysis teaching materials.
- The Sunlight Foundation's guide for using NodeXL for Twitter analysis.
- NodeXL YouTube tutorials - part one looks at visualizing a simple workbook, part two covers importing and analyzing data, and part three explores more advanced techniques.
- Penn State University's online guide provides an in-depth walkthrough of NodeXL, as well as Flickr, Facebook, Twitter, YouTube, and hyperlink specific analysis tutorials.
Download NodeXL here.
Download NodeXL Pro here.
Visit the NodeXL Graph Gallery here.