Skip to content
thedod edited this page Apr 13, 2013 · 26 revisions

Introduction

In a single sentence: CableWeaver uses D3 visualization to let you interactively inspect Cablegate reference clusters generated with Cable2Graph.

For a longer introduction - see what is CableWeaver.

As you can see on the sidebar, there are 3 stages involved:

  1. Search for graphs
  2. Select a graph
  3. Inspect the graph
  4. Share what you find

1. Searching for graphs

In the 1. Search for graphs section, enter the MRNs (Message Reference Number, e.g. 10TELAVIV413) you're interested in. Simply put them in the text box, one MRN per line.

You can search for MRNs in any of the following fine establishments.

If you don't have any specific MRNs in mind, and just want to get the feel of the system, click on the example link and it would paste a few MRNs into the text box.

2. Selecting a graph

In the 2. Select a graph section, you'll see a menu of all graphs that contain any of the MRNs you've entered in your search.

Note:

  • Not all MRNs have a graph containing them. If you get a No graphs found message, try other MRNs. Note that searching for a large number of MRNs is not a costly operation, so you can start by entering any MRN that might be relevant to what you're looking for, and see which MRNs actually appear in CableWeaver's graphs.
  • Graphs are represented by the MRNs they contain, and you may see more than one entry with the same "name", but they are different graphs that happen to contain the same MRNs from your search.
  • Some graphs contain too many nodes to be useful (at least with CableWeaver), but in such cases you should also see "manageable" partial graphs that contain some of the MRNs you're looking for.

3. Inspecting a graph

Once you select a graph, you'll see it on the main area of the browser, as a collection of nodes (circles) representing cables and links (lines) representing reference between 2 cables. Each link has an arrow-head pointing to the referred-to (i.e. earlier) cable.

Navigation and manual layout

  • Hover over a node to see details of the cable it represents.
  • Click on a node to open the related cable in a new tab.
  • Ctrl+click on a node to toggle whether it's selected (selected nodes appear larger). This also affects the timeline display (see below).
  • Shift+click on the node to toggle between manual placement and auto placement modes (see below). Default is auto placement.
  • When you drag a node, all nodes in auto placement mode rearrange themselves as if the links were "rubber bands".
  • When you release a node that is in auto placement mode, it snaps into place according to the "rubber bands" attached to it. This is why it is usually best to switch a node to manual placement before dragging it.

Legend

(also for exported graphs, e.g. at the CableWeaving archive)

  • Larger nodes represent selected MRNs (initially - the MRNs you were searching for are selected, but selection can be toggled with Ctrl-click on the node or on the corresponding timeline link).
  • Node's fill color represents sending Embassy.
  • Border is dashed if cable is missing, otherwise the color represents classification (UNCLASSIFIED, CONFIDENTIAL, and SECRET).
  • Border is darker when node is in manual placement mode (not applicable for exported graphs).
  • Border width represents node's betweenness centrality (in how many shortest-paths between pairs of cables it appears).
  • Link arrowhead points to the referenced (i.e. earlier) cable.
  • Link width represents betweenness centrality of the reference.

Timeline

  • All cables represented in the graph are also sorted by date in the timeline.
  • Selected cables (the ones that are represented by a larger circle) also have a pushpin near them.
  • Ctrl-click toggles selection. This also affects the corresponding graph node, so toggling a couple of times is a way to locate it on the graph.

Sharing your findings

Once you select a specific graph, you'll see the share and export buttons.

  • The share button lets you copy a link that would open CableWeaver at the same graph you're watching now, with the same set of selected cables (also called "story line"). The graph wouldn't look exactly like the one you see now (because the layout is dynamic), but the graph will be isomorphic to it (i.e. it would be possible to edit the layout manually and make it look exactly like "your" graph).

  • The export button "freezes" the graph's layout and lets you copy HTML of an entire web page containing a static graph (where selected nodes appear numbered), followed by a list of all selected cables, each with general details (title/date/etc.) and a link to the cable's page. Here's an example of the result of such an export.

    You can either copy this HTML nad paste it into a text editor, or save it as a Github gist.

    In both cases, you can add a summary, quotations from cables, remarks, etc. (see examples at CableWeaving).

See also: