For this blog post I decided to look into “Using Palladio to visualize ads”. This post is about visualizing runaway and fugitive slave movement patterns using fugitive ads. If you would like to look at the original project it is found here. The main argument is those in more southern states would tend to escape south as opposed to north. The writer use Palladio as the tool for creating the visuals which provides an easy way to create graph based visualizations. Each node in the graph represents a location, either the starting location given by the ad owners address, or the end location given by the location of arrest. The edges in the visualization represent the presumed path of the escapee. An example is given below.
There are a few problems I have with the visualization, one it does not show directionality. The edges should be directed edges but Palladio does not give this functionality. This not ideal as the argument is based on the migration pattern of the fugitives which is directionally dependent. We also are not able to see if a node is a start or destination node, which again hurts the writers argument.
Another problem is how the graphics are displayed in the post. They are still images which display each area and graph separately. When trying to show multiple graphs at a time the author simply incorporates multiple images into a single figure. I believe that an interactive viewer that shows all graphs together would be much more interesting and intuitive for the user.
One interesting components these visualizations is it combines a graph based visualization with spacial context. The graph is displayed over a map which allows you to see the patterns along with its geographic location. This is an important component of the graph as location is important to the authors claims.
Overall the downfalls of the visualizations are due to the restrictive nature of Palladio. I would be curious why this software was chosen over alternatives which could enhance the visuals. With features such as directional edges and named nodes, the connection between the authors argument and the visualizations would be much clearer.
