one person, one dot (applications)


In a previous post we have examined the usefulness of a dot map and its potential limits for social analysis. The aim here is to dig a little bit deeper into some of its applications beyond ethnic segregation. In the original post we have mentioned that this methodology could be easily adapted to other social phenomena. Here we explore how it has been done, using some illustrations based on two key areas: vote and demography.


one dot, one vote


In the moment that it becomes possible to associate the geographic location of a person in a map, it also becomes evident the potential application of this method for electoral studies. The existence of zones in which dots of a same color are concentrated (voters or votes in one party) reveals electoral enclaves that, faced to other demographic characteristics, highlights possible ecological conditioners of voting behavior. Hegemonic or disputed areas, nationalization or regionalization of party voting, spatially dense or scarce areas, these characteristics affect both voter and party expectations, as well as are potential conditioners of relative costs of an electoral campaign or the provision of social services.

We illustrate these applications through the analysis of the Congressional Dot Map containing the vote to the U.S. congress on the presidential elections in 2008. Following the same structure of the Racial Dot Map, this dot map allows user to identify where the vote in both Democrat and Republican parties is clustered, as well as to spot disputed areas.

Fig. 1. Congress Dot Map, U.S. Presidential Elections, 2008 (2013)

PR2 - Congress

(Click on the image)

This U.S. Congress map is well achieved, since it presents two characteristics that add value to the map by situating dots into specific contexts that help to interpret the results displayed. Firstly, it allows users to overlay electoral district boundaries to dots, revealing if voting patterns in each district is hegemonic or contested. Secondly, it allows the visualization of the ethnic distribution within each electoral district, enhancing its usefulness in order to understand the relationship between demography and voting behavior in the U.S.

The demography of vote in the U.S. is a key issue in the organization of electoral disputes. Differently from other countries, in the U.S. it is common practice to change electoral district boundaries for electoral purposes. This practice, known as gerrymandering, constitutes a strategic resource for both parties. The manipulation of electoral boundaries according to previous voting patterns and demographic characteristics of residents can be key for the victory of a party in these single member districts. It is precisely this aspect that converts this kind of visualization into a particularly useful tool for revealing the logic behind the administrative organization of elections.

Despite of its evident utility, this map gets just half the way in terms of its knowledge generation potential. With a little more work, this visualization could include some aggregate information, such as the total number and percentage of votes in each party or the degree of spatial concentration or density of voters. Such data could help developing hypotheses and exploring patterns between the vote in a certain party and other contextual aspects of districts.


one dot, one job


Another application of this method can be observed in “Where are the Jobs? Employment in America 2010”. Using data of the Longitudinal Employer Household Dynamics of the U.S. Census Bureau, Robert Manduca, a PhD student in Sociology from Harvard, has categorized jobs according to four major economic sectors: (a) Manufacturing and Logistics, (b) Professional Services, (c) Healthcare, Education, and Government, and (d) Retail, Hospitality, and Other Services.


Fig. 2 “Where are the Jobs?”  project

PR2 - Jobs

(Click on the image)

As the maps of ethnic segregation or voting patterns, this visualization reveals the spatial concentration of jobs according to economic sector, suggesting an association between ethnicity, income, voting behavior and the economic sector in which individuals are employed. In particular, it draws the attention on how the economic structure in each city can be represented in the spatial distribution of jobs, as well as in the patterns of employment concentration of professional services, healthcare, education and government that are clustered in more central areas, while manufacturing are more peripheral or suburban in major cities.

Spatial segregation based on economic sectors can reveal important divisions in terms of class or status within the urban space. Other maps, such as the one from City Science covering Melbourne or another published by the New Zeland Herald go in this same direction, enabling the visualization the places where rich people live and the degree of spatial segregation in terms of class.

This kind of chart, besides identifying richer and poorer zones, provides us with an instantaneous picture of the economic organization of a city as a whole. If it is combined with other demographic categories such as age, ethnicity or occupation, it would allow us to establish potential associations between demographic characteristics and the incidence of poverty or inequality in space. A straightforward illustration of this can be found for the case of Melbourne, where the area between Southbank and Carlon is characterized by the predominance poor and young Asian immigrants.


One step forward: demographic data explorers


As we could see before, the examples of application of dot maps using demographic data are various and include other information on households and immigration. Nonetheless, their predominantly experimental and unidimensional character (just one variable at a time) limits its capacity to serve as tools for discovery and exploration and, therefore, restricts their application in other fields, such as business and potential scientific applications.

We would like to mention here some experiences that go one step further towards more developed applications in terms of data exploration. Such examples stand as pioneers and present all the potential to become references for the visualization of multidimensional demographic databases in the near future. The core of these portals can be found in the fact that they facilitate the simultaneous exploration of multiple variables using comparative map panels.

Two examples deserve special mention: the Social Explorer, from the company with the same name, and the Synthetic Microdata Household Viewer from RTI. Their relevance derives from their usefulness as tools, not so much by the content they share. Social Explorer is probably the best due to its functionalities and the available data covered (census and surveys from 1790 until the present). The dimensions covered (in different scales from states to blocks) include data on households, income, house costs, labor market, crime, carbon emissions, among many others.

Fig. 3. “Social Explorer” Project (2007-2016)

PR2 -Social Explorer

(Click on the image)

The potential use of this type of tool is huge, since they constitute platforms deliberately conceived for the comparison and exploratory data analysis in different scale resolutions. Their main limitation, though, is on the exclusive use of maps for data analysis, not including other types of visualization. The inclusion of charts such as scatterplots, histograms, radar charts and correlation plots would convert such applications in extremely useful instruments for research and knowledge development. Beyond mere visualization portals, they could offer useful platforms for scholars, policymakers, political parties and companies.