Skip to end of metadata
Go to start of metadata
Description

Our version of GUESS includes the ability to save and restore node positions via new File menu options. To save node positions, choose File > Export Node Positions. To restore node positions, choose File > Import Node Positions.

Node position files are actually just standard CSV files, so we recommend that you choose a .csv file extension when saving node positions. If you examine a saved node position file in Excel (or any other software capable of reading .csv _ files), you will see that there are three columns: originallabel, x, and y. originallabel corresponds to the _originallabel attribute on nodes, and this is how it matches which node to give the specified x and y coordinates to when restoring node positions.

You might also notice the first two entries in this node position file: [cameraScale] and [cameraTranslate]. As mentioned in the below tutorial, one common use of this Node Locking feature is for laying out time sliced networks (or extracted networks from time sliced data). The _[cameraScale] and _[cameraTranslate] are useful for this because they also specify the camera position and zoom level used in a layout.

Workflows

One might wish to use GUESS to visualize networks of a time sliced nature. Here is a workflow for how to do this in Sci2:

  1. Load the initial dataset (the full dataset) and do the desired network extraction on it (the full network). Load this full network into GUESS, lay it out as desired, and then save out the node positions.
  2. Back in Sci2, choose the full dataset and then the Preprocessing > Temporal > Slice Table by Time algorithm.
  3. You can then choose the resulting time sliced tables and do the same network extraction on them as on the full dataset.
  4. You can then select these networks, load them into GUESS, and restore the node positions you saved out in step 1.

Although there is potential for further automation, the Node Locking feature in GUESS greatly reduces the overhead of laying out "time sliced" networks.

  • No labels