Wednesday, May 7, 2014

Network Mapping and the Civic Club of Philadelphia



The Civic Club of Philadelphia:
Women on School Boards Committee, Nominees and Elections, 1908-1911


For the past few months I have been playing around with Gephi, an open source graphic networking tool in order to map the Civic Club of Philadelphia's attempts to place women on ward school boards in Philadelphia. The visualization I have ended up with displays the Civic Club nominations, the women who were elected, the wards they lived in, and the year(s) the events took place. This post primarily focuses on the process of making the graph and speculates about the ways scholars, students, or any interested party could use it.

Beginning this project, I spent an enormous amount of time organizing data. The club's campaigning for women on school boards started in 1895. For that year they published a great account of the entire campaign. Unfortunately, beyond 1895 I had to rely on often incomplete newspaper announcements of the club's activity.  In the end, I focused on the years 1908 to 1911 because the Civic Club Bulletin (first printed in late 1907) published complete lists of Civic Club nominees.

(Note: The Civic Club of Philadelphia Records are held at the Historic Society of Pennsylvania. The archives do not have a finding aid, but were well organized internally before the club donated them. Listings of nominations are in the Newspaper Clippings binders and the Civic Club Bulletins are compiled in volumes. Link to records page.)



Master Excel: Women on School Boards, with notes and sources


Gephi graphs are created out of nodes (the dots) and edges (the lines connecting the dots). At first, I was not sure if this type of network mapping would work for my data. The women would never be connected directly to other women, the wards would connect with the same people, and how to visualize both the women and their status in the campaign was a mystery. I eventually decided to make nodes out of women, wards, and events. Each node has a numeric Id and a label. The edge list connects the events to women, and the women to wards though labeling sources and targets.



Edge CSV for Gephi



The fun started once I uploaded the data to Gephi. The layout and color potions are addicting to play with and can create some amazing images. Nevertheless, an image like the one below, however awesome, is completely useless for my purposes. Figuring out how to format the graph so that it could actually tell me something required a combination of historic and aesthetic thinking that I enjoyed employing, especially since usually I am entirely one or the other. 




Playing with Formats in Gephi



The layout I decided on was to locate the ward nodes in their approximate location on this historic map of wards in Philadelphia from 1914. This would add an extra dimension to the graph and place the club's activity in a spacial context. 



Philadelphia Ward Map, 1914 (original here)



The resulting layout dispersed the wards, and an additional layout consolidated the women around where they lived. I manually dragged out the nominations to the left, and the elections to the right. Had I been thinking more ahead I would have also located these outside of the perimeter of the map.

Wards Locked and Formatting almost Finished



With the layout decided on, I still had to figure out the color scheme. Like layouts, pretty colors distract me, and finding a scheme that works with the data was a challenge.



Cool Looking, Though Not Very Useful, Graphic Display



Eventually, I decided on a red to blue scheme, where red is the earliest year, 1908, and blue is the latest, 1911. The nominations are desaturated and the elections are fully saturated. Ideally, this allows viewers to more easily see who was nominated, elected, where and when. Though still a little cluttered, especially around center city, the information is much clearer in the final graph.



Final Graphic (minus map)



This final graph, and the overlay with the ward map at the top of this post, provides a way to understand the Civic Club's activity and local school election that is difficult with the lists of women published in newspapers and bulletins. As a visual learner and thinker, I often feel a little left out of the mainstream historic discipline because it is so partial to strictly written sources and written interpretations. Visualizations can force people to approach their sources in a different way than they would if they were writing.

I am still working with how to interpret and present the graph. Currently, it serves as a good platform to ask more questions about the topic and to generate interest in a largely unknown movement.