Visualize network on physical map

[glstott]

Build python script to map network on top of a map. 3 stages: connect to neo from python, generate network in python, and map nodes to locations (zip code blocks with jittering

[glstott]

• Mapping of nodes to locations with zip codes complete. I used a quick python script to parse the coarse zip code (some use the 9 digit instead of 5 digit zip) and find the centroids for each zip, then send those to a CSV file for insertion into Neo4j.
• Gephi quite readily maps nodes to their geographic coordinates through a plugin called GeoLayout. Then, select the variables in the layout tab.
• Cytoscape can do something similar via enabling the x and y coordinates in style properties, then mapping each to the needed variable found in the node. Cytoscape's is a bit buggier, and will require adjusting of scale and other style properties to match.
• There is no clear way (which I can find) for either of them to place a map beneath the nodes without a great deal of manual scaling and centering.

[glstott]

• Created draft python script to calculate within-zip sample counts for a list of parameters. This will allow for a bubble-view of geolocated data.
• Generated cytoscape view for bubble-view of geolocated data. For each zip, there are 5 time sliced nodes and 5 time sliced edges.
• Attempted to create pie chart visualizations for each node. Difficult to set up, then when I thought it was set up properly, it repeatedly caused crashes. It may be due to my variable choice, pango_lineage, which has 81 possible slices. A better choice would have been clade.

[glstott]

• Attempted again with clade since it has just 8 categories. Again, encountering memory issues. May be due to concurrent processes, but regardless, if it can't handle visualization with 8 GB RAM, it probably isn't a viable option.
• Set up the visualization to ignore that aspect for now.

[glstott]

Bubbles geolocated, no additional visualizations on top of that view. I didn't find this view particularly useful.