Blog

Presenting “Mapping Consumers” at a World History conference

Two weeks ago, I went to Northeastern University’s 10th annual Graduate World History conference. My paper was largely based on the sources that I draw on for my CHI project. This was my first time showing my site to the public, and the deadline for the conference presentation was a great motivation to get my site up and mostly functional. This is what my site looks like currently:

Mapping consumers home page

One of the things that was helpful about presenting this to an audience of strangers was learning which parts of my site are self-explanatory, and which need more context. I found myself having to explain the colourful marker clusters, which is a feature built into the Bootleaf template that I’m using. When there are more than a certain number of markers in a certain area at a certain zoom level, it will create a coloured cluster representing this group of markers. However, I realized while presenting that these clusters somewhat obscure the argument I want to make with this map — that testimonial-writers in the black press were from rural places all across the South African countryside.

Based on this experience, I’ve now changed the zoom level at which the marker-cluster feature works. I’ve turned off the marker-cluster feature at a lower zoom level, so now as you zoom in, you’re able to see the spread of testimonial-writers across the map. Like so:

Markers showing testimonial writers in the Eastern Cape region

Another positive take-away from this conference was that because it was a World History conference, I was able to think in new ways about how my local South African project is contextualized in global history. My paper made the argument that testimonials show how local South Africans responded to global marketing campaigns (in particular the well-funded campaign for tea consumption put on by the Tea Market Expansion Bureau).

Mapping Movement

The most rewarding aspect of my project has been working on my map since it has been the most challenging. I opted to use leaflet.js to build my map, which shows how many migrants each region in Russia sent to Moscow in various years between 1970 and 2000.

I first used leaflet.js during one of our development challenges last semester, and I found it a bit confusing at first. The leaflet.js website has a tutorial, but it was unclear where to place the information for the JavaScript and CSS sheets with the other information from the bootstrap that we used. If we failed to properly place this information, the map would not run. Once we figured out where to place and how to properly label this information, the map showed up.

One of the major issues I faced was not inherently technical but dealt with the best way to integrate my dissertation research into the map. Soviet authorities sometimes changed how they recorded statistical information. Every year through 1985, the Moscow Committee for Statistics (Moskomstat) recorded the place of origin by region (oblast) for migrants who arrived in Moscow from other locations within the Russian Soviet Federative Socialist Republic (RSFSR). For Soviet citizens from one of the other 14 republics, only the republic, and not a specific region within the republic, was noted. Approximately every 5 years, Moskomstat recorded that nationality of all arrivals, which complicated my understanding of who these migrants were. Although the number of migrants from other Soviet republics began to rise throughout the 1980s, migrants remained primarily ethnic Russian, meaning that Russians were leaving other locations in the Soviet Union for the RSFSR.

Political reforms during perestroika and the breakup of the Soviet Union also changed statistical recordings. During perestroika, Moskomstat often simply recorded various types of internal migration: intra-region, inter-region, and inter-republic without specifying places of origin. In the post-Soviet period, the Northwestern Region separated into the Northwestern and Northern Regions, and across the Russian Federation, several regions changed their names. On my map, I have pop-ups for each small region surrounding Moscow (other areas in the Central Region), each large region across the RSFSR, and each republic of the former Soviet Union. I am currently deciding if I should include any other additional information in my popups. My next goal is to find an appropriate Soviet-era map, printed in English. The current map tiles that I am using through leaflet.js use local languages, which might make the maps difficult for my English-reading audience.

Networked Society’s Celebrity Networks

I have had to tone down some of my initial goals for what I thought my network would be, but am still going to endeavor to make a semi-interactive network and a website to house information for people who are interested in viewing the affiliations that extreme right celebrities have.

However, considering the role that the Internet has had in the resurgence of populist fascism in the U.S. and beyond, I felt that paying special attention to “celebrities” with significant online presences might be an interesting way to visualize these connections. Although the term “network society” was coined in the early 90s, the origins of it are from before that in terms of how society would eventually be connected by wired communications. Famously, Castell’s work The Rise of the Network Society laid down the foundations of how he envisioned society would eventually function: it is not that networks merely exist within our societies, but rather are the very basic foundations of them. Although of course, I’m not saying that the Internet is to blame for the rise of extreme right-wing politics, it did play a part in the accessibility of the ideology to millions of people.

In that regard, how do we envision networks of Internet celebrities within the networked society? What conditions led to their rise, and how might these conditions also lead to their downfalls? These aren’t necessarily questions that I’m grappling with since my project will be mostly descriptive and for visualization purposes, but are ones that can be dealt with later in academic work. In terms of downfall, we have seen how although the affordances of the networked society have been great, they have also been detrimental: websites are being shut down, subreddits are being removed, and Twitter accounts are being banned. But this brings about another age-old issue in terms of how digital space is governed and regulated, which are things that will constantly be debated as technology outpaces the policies that control them. This project, although not necessarily intended to, has really helped me think about the power of platforms, infrastructures, and their roles in shaping and inhibiting social movements – which is the basis of my dissertation, stay tuned for more.

Theoretical musings aside, I’m following a tutorial on how to make my network graph because a lot of network mapping software does not easily allow for the creation of event-affiliation or bipartite networks. However, software like R that requires a lot of manual coding do allow for the creation of these visualizations. Although seemingly very descriptive, event-affiliation and bipartite graphs can be helpful in determining who belongs to some organizations, thus increasing their likelihood of encountering people who otherwise may be disconnected. I’m making slow progress on the graph because it is in d3.js, but there is an R tutorial that may help me move along faster. Either way, I’m grateful for the experience because instead of pointing and clicking, I’m actually managing to teach myself the foundational code of what makes up these networks. Foundations of foundations, we’ve gone meta.

Tutorials for the curious:
How to Make an Interactive Network Visualization
Network Analysis and Visualization with R and igraph

Charts, Graphs, and Coding, Oh my!

For my remaining blogs, I plan to address the different technological elements on my website. First up are the graphs that I have created using AM Charts, a website that allows you to build JavaScript charts, graphs, and maps. For my website, I built graphs that show for Moscow: births and deaths, natural and migration-related population growth, total population growth, and changing nationality and place of origin.

AM Charts provides a very user-friendly system that provides you with multiple options for building your charts and graphs. First, you can download free versions of AM Chart’s products. The website advertises that, “The only limitation of the free version is that a small link to this web site will be displayed in the top left corner of your charts.” The download comes with over 80 sample charts that you can view and examine the code. The graphs range from simple bar graphs to complex bar and line graph combos with interactive elements that explain data points. The download also comes with a brief tutorial and various JavaScript and CSS sheets. The second option for using AM Charts is paying for it, which I cannot comment on, except to say that if the creators are correct, it should be similar to what I have described.

I have also used the third option: the online chart editor. This option requires you to create a free account with AM Charts. The online chart editor provides you with several dozen templates to work with. After selecting your template, you may either upload a CSV file with your data or input it manually. I found it somewhat difficult at times to adjust the legend and axis labels, but I discovered that there are multiple ways to address these issues. The chart builder allows you to select and alter options through an input interface, but you can also access the code to edit your chart there. Once your chart is complete, you can use an embedded link to present it on your website. AM Charts will save your graphs that you built while logged in to your free account.

The graphs that I have created have helped me understand my research data better. On my website, I have three line graphs that depict population growth in Moscow. The first shows births and deaths on the same graph. Until 1984, births and deaths rose together, which did not radically alter Moscow’s population. In 1989, deaths surpassed births and continued to do so throughout the 1990s. My second line graph illustrates population growth related to migration and live births on the same plane. Migration consistently played a larger role in contributing to population growth, but tended to mimic the patterns of natural births. In 1989, deaths surpassed births while migration did not contribute to Moscow’s population decline until 1992. Migration to Moscow dropped sharply, but rebounded before birthrates did, contributing to population growth. The last graph shows Moscow’s overall population growth.

My current dilemma is how to use graphs when methods of recording data changed. In the Soviet period, place of origin and nationality were not necessarily the same thing. For example, an Armenian could have lived in Azerbaijan before moving to Moscow. Soviet statistic takers in Moscow recorded place of origin for all migrants annually, and nationality every several years. Statistics in the post-Soviet period tend to only record place of origin, which obscures information and poses a challenge in creating graphs.

Project update: refining my map

In the past month, after several weeks of despairing tinkering, I finally got the features in my GeoJSON files to appear as points on my map. I now feel much more comfortable playing around with Javascript to see what my map can do. Currently my map looks like this:

Now that the basics of the map are there — markers for each testimonial advertisement in a newspaper, clusters for concentrated areas of markers, and a sidebar with click-able information — I want to add more complexity to the map to represent more information about testimonials and the history of consumer culture in South Africa. One of the things I’m interested in is highlighting differences between urban and rural writers of testimonials. Since I haven’t been able to find an appropriate historical map of South Africa from the time period, Ethan Watrall gave me the idea to create different-coloured markers for urban or rural locations. Now, I’m working on this task, using population information from the 1936 census. My goal is to represent urban or rural status using different coloured markers, with different shapes representing the different newspapers that the information comes from.

Adding this additional complexity to my map will force me to consider exactly how to define “urban” and “rural”. One metric I’m considering is whether or not a town is listed in the 1920 census list of “towns and villages possessing some form of urban local government.” But this is perhaps a problematic category, as the structures of “urban local government” might have been granted differently depending on the racial make-up of a locale. In that case, my best option may be to pick a certain number that I consider to represent the border between urban and rural, and simply trust that the 1936 census gives a fairly accurate population count.