Connecting Occupy

This project uses data from the Occupy Wall Street Project List (Issue 2, April-May 2012) to map connections between the Occupy movement and other organizations. The data for this small study comes from the project descriptions, which often include lists of partners.

Using the open-source network analysis tool Gephi, we can examine these connections in various displays. These graphs provide an abstract spatial sense of how groups are related, as well as the activity and flow across this structure. By coding partners by type, we can also see how certain types of groups (e.g., NYCGA working groups) are connected with other types (e.g., unions, religious/spiritual organizations, political parties).

Figure 1. Project partners in a force-directed network graph. Each group is represented as a node connected to other groups (nodes) through a project (line/edge). Some nodes are bridge nodes, connecting different parts of the network. In some of the tighter clusters, organizations (nodes) are close to each other because they partner with each other on the same projects. Some nodes (esp. yellow NYCGA nodes) are prominent in the sense that they have connections to many other nodes.

NYC General Assembly (yellow) | Larger Occupy movement (red) | Political Organizations (green) | Religious/Spiritual Organizations (blue) | Unions (purple)

 

Figure 2. Project partners in a force-directed network graph (curved lines). Each group is represented as a node connected to other groups (nodes) through a project (line/edge). Groups cluster together when they partner on the same project (e.g. groups working primarily on “Health Care for the 99%” are largely clustered at the right of this graph; yoga, wellness, and spiritual groups are clustered at the bottom right). As groups partner with multiple projects, they are pulled away from the cluster surrounding any one project, and pushed toward the center of the graph (e.g., the OWS Arts & Culture working group near the center of this graph has partnered with many different projects).

NYC General Assembly (yellow) | Larger Occupy movement (red) | Political Organizations (green) | Religious/Spiritual Organizations (blue) | Unions (purple)


Figure 3. Project partners in a radial graph with edges colored by target node type. This graph shows that most connections terminate in a wide variety of groups, including larger Occupy movement (red), political (green), religious/spiritual (blue) and union (purple) groups—and especially the latter three given how few of these groups are included in the overall dataset. Since the partnerships contained in the project list are neutral with respect to order (A with B = B with A), this graph does not show the direction of activity per se, merely the prominence of the various types of organizations that are connected by common projects.

NYC General Assembly (yellow) | Larger Occupy movement (red) | Political Organizations (green) | Religious/Spiritual Organizations (blue) | Unions (purple)

Next steps:

The general method developed here may be extended in several ways:

  • Additional data sets—this “parternship” framework could be extended to other datasets, including event planning and endorsements
  • Geospatial data—a number of these groups focus on specific locations or neighborhoods, and a geospatial layout would show the way in which this network structure operates over physical space
  • Temporal data—by adding other issues of the project list or timestamped data, we could observe this network grow and change over time

3 Comments

  1. Hi

    I’ve been working on something extremely similar for the past nine months as a project of OWS. In fact, we collected the data you are using. It would be great to talk further and see what kinds of overlaps we have. This research and its findings are supposed to go back to the OWS movement in some version soon — that was the arrangement we reached in through working with a couple of different OWS working groups as far back as December 2011. I’ve sent out invites to collaborate on this project in the past and was unaware anyone else was also working on this. It would be awesome to get together and talk!

    Please get in touch me! I’ll definitely be at the hackathon on Friday. I have built out an extensive version of the data with a variety of properties not part of the newsletter you used in the example above.

    Please email me jeo-tech@earthlink.net.

    Solidarity,

    James Owens

    • srt0

      You should post it James. The more work on this site, the better!

  2. Pingback: #OccupyData Hackathon May 18-19 | #OccupyData NYC