Opening its doors in 1887, Pratt Institute has long been a physical and cultural presence in the Clinton Hill/Fort Greene neighborhood of Brooklyn. Erected in 1907, the Pratt Townhouses have weathered over a century of campus changes and development, consistently providing stable and relatively affordable housing for members of its community and external parties alike. To help investigate the history of the institution, I sought to provide researchers and knowledge-seekers with another tool to conceptualize and better understand the lives of the educators, administrators, and working people whose efforts have been essential to the school’s operation since its founding. To this end, I constructed a dashboard containing a map of the Townhouses, a timeline of the residents at each location, as well as a Sunburst chart giving the duration of their overall tenancy to promote exploration and provide opportunities to learn about the Pratt community members and adjacent individuals who lived in these spaces.
Audience and Approach
When planning my initial approach to this project, I gave careful consideration to whom the most likely user would be of historical data related to residents of the Pratt Townhouses. The first group that came to mind is that of researchers interested in exploring the early dynamic of the Clinton Hill neighborhood, investigating the development of early creative communities, and developing a more comprehensive understanding of the lives of the artists and educators. Within this set of researchers, I suspected that there would be a large contingent of current Pratt Institute staff and faculty looking to deepen their own knowledge of the school’s history or to inform their own academic pursuits. This particular user group and line of inquiry led me towards an exploratory model beginning with a “One Big Map” structure. Additionally, I could foresee a secondary audience of current students or alumni with a personal curiosity in who may have previously lived in the same spaces that they recently occupied or currently reside. To add informational depth as well as additional means to engage with the data, I elected to incorporate additional visualizations, ultimately constructing a dashboard.
Data
The original source for much of the data used for this project came from the archival holdings of the Pratt Institute Archives. The addresses, lease-holders, and rent amounts were compiled by a previous researcher, Patrick Kingchatchaval, from residential ledgers and leases generated by the Real Estate Department of Pratt Institute. To expand the available information on the residents, I consulted with Pratt Institute course catalogs from 1907 to 1971, which frequently listed their role at Pratt Institute (faculty-member, staff member or administrator, etc.) as well as identified the Department or Office that they were associated with during their employment. For more recent faculty and staff, I searched through Pratt Institute publications such as the Prattler, the PI Newsletter, and the Prattonia (the yearbook since 1921), whose reporting often included Pratt staff who otherwise would not have been in the Course catalogs, such as Buildings and Grounds staff and administrators.
After an initial round of research through Pratt publications, I noticed multiple gaps and “unknowns” when it came to the occupation and status of many of the residents. Attempting to create as complete a dataset as possible to facilitate more in-depth inquiry into individuals, I expanded my research by consulting the United States Census. Publicly available from 1900 to 1950, this resource was invaluable in identifying biographical data for those residents of the Pratt Townhouses who were not Pratt employees or otherwise associated with Pratt. This additional information provided an interesting perspective, revealing a noticeable trend in the professions of those who lived at the Townhouses, but were not themselves explicitly attached to the school.
Design/Layout
Informed by existing archives and special collections visualization efforts (in particular Princeton University’s “Princeton & Slavery” project, I chose to construct a dashboard model with three primary elements:
1. Interactive Map of the Townhouses
2. Gantt Chart
3. Sunburst Diagram
Map View
The map is the primary focus of the overall dashboard which is reinforced by the amount of real estate that it occupies within the overall layout. Built by connecting the Townhouses’ individual addresses with a shapefile provided by the New York City’s Primary Land Use Tax Lot Output (PLUTO) database, each clickable point on the map corresponds to a specific address within the Townhouses. Viewing these points together on the map highlights the close proximity of the buildings with one another as well as situates them within the surrounding neighborhood.
There are two primary means of navigation/interaction with the map, selecting the individual building to filter to an address, or selecting a single year from a dropdown menu to allow users to see who occupied the buildings at the same time. Initially, users had the option to select the addresses from a dropdown as well as by resident, however, user feedback indicated that these options were a bit duplicative (in the case of the address) and overwhelming (in the case of the Residents’ List).
The tooltip is fairly minimal, providing users with the address of the specific building as well as the number of lease-holders overall from 1907 to 1971, when no filters have been engaged. If users use the floating “Year” filter, the tooltip will populate with the name of the specific lease-holder of that address from that Year.
Gantt Chart – “Resident Timeline”
In addition to the map, I included a visualization that I thought that highlighted the individuals’ experience as well as reflected the continuity of residency at the various Townhouse locations. Sorted by earliest tenants in 1907 up until the more recent from 1971, users can trace the succession of residents from left to right and in descending order. While the primary filtering mechanism is intended to be the map, users can select a specific address to view only that building or individual residents to highlight them further. Selecting either an address or individual tenant filters other visualizations within the dashboard to provide additional interactivity between elements in the dashboard, as recommended by reviewers.
Sunburst Diagram – “Year at Address for Residents”
The final visual element of the dashboard is a Sunburst diagram, consisting of an inner ring with the 27 individual buildings and an outer ring containing all of the residents of the buildings with different slices corresponding to the number of years that they were residents within the Pratt Townhouses. I was drawn to the Sunburst diagram as a visually interesting element which displayed similar information as visitors might glean from a bar chart, while not being too duplicative of the Resident Timeline. By providing a total count of years of residents, it reproduces data within the timeline but in a more direct and understandable manner.
Archival Records and Images
I introduced one additional interactive element to the map which is a series of links to archival records or images related to the buildings and residents of the Townhouses. When users select an address, they will see a link to “View a Historic Image of Building Group” which will take them to images from the 1940s of that row of buildings. The current images linked are from the 1940s WPA-New York City Tax Department photographic documentation project, whose prints are currently held by the Municipal Archives. Incorporating links to images in the future adds additional depth and visual engagement to the dashboard as well as further connects the information to its source, offering an opportunity for further exploration.
Evaluation and Reflection
To get feedback regarding the visual arrangement and experience, I reached out to colleagues who were familiar with the source material and had participated in similar mapping projects. With this background, I believe that they were good stand-ins for my imagined primary audience of expert researchers. Speaking with them in person after they had time to explore the dashboard on their own, they expressed an interest in increased interactivity between the elements, a more simplistic and less “busy” dashboard, and increased availability of information related to individual residents. This feedback aligned with recommendations and suggestions that I had received from the professor as well as my classmates during review sessions. In response, I removed extraneous filters, updated the dataset with more in-depth individual information gleaned from expanded research , and incorporated additional information into existing tooltips.
One interesting piece of feedback was fairly widespread support for the Sunburst diagram. This enthusiasm ultimately led to my keeping it in the dashboard despite a lot of difficulty with its functionality. I find that it is very helpful once engaged, but when no filters have been selected, the effect can be overwhelming visually due to the wide color palette (27 unique colors for the addresses and 174 unique residents). At this time, I have not been successfully able to engineer a solution that hides the diagram when “no-filter-is-in-place” while not breaking filtering functionality elsewhere in the dashboard. I decided to leave it in for this iteration while considering a replacement visualization (see “Future Developments”) containing similar content.
User Orientation Help
Another consistent point of feedback from users was the need for some orientation to the functionality of the map. Individuals were eventually able to navigate and explore, but there was an initial learning curve that could be discouraging to those unfamiliar with the platform. At the suggestion of a classmate, I created a set of instructional slides which I linked to an image object on the dashboard labeled “Need help with the map?” It is positioned at the bottom of the map in blue to make it distinct from other visual elements nearby, and clicking on the button takes the user to the slides in another tab.
Initial Findings
Spending time with the data and interacting with the visualizations provided interesting insights and trends regarding the people who lived in the Pratt Townhouses from 1907 to 1971. I was previously under the impression that it was explicitly Pratt staff, students, or faculty with a few notable exceptions, which is not the case. It seems that a good number of individuals living in the Townhouses worked in the education field as public and private teachers at both the high and middle school levels. While I did not ultimately incorporate it into the dashboard, I developed multiple visualizations including a pie graph and treemap, which reflect the proportion of Pratt associated individuals versus those with “external” affiliations as well as the Pratt department or private company that employs them. With additional work such as larger categorical grouping, these visualizations could be included into another dashboard to supplement the current interactive one.
Another interesting point of information surfaced by the visualization is that of the length of time that many individuals lived at one address within the Townhouses. There are many cases where individuals like Isabel Ely Lord, Beulah Stevenson, and others lived there for 35 years or more. Some interesting social trends were reflected in the data as well, such as the practice of allowing family members of deceased residents to continue living at that location by transferring the lease over to their name. A few individuals who took over the lease of a spouse or relative such as Ellen Bivins, Bertha Edminister, or Oona Burke may in fact have lived there for much longer than is accounted for by the visualizations, as a result of patriarchal practices regarding who is the head of household and “lease-holder.” With this knowledge, the timeline really highlights their duration of residency in a way that the Sunburst diagram does not.
Future Developments
After additional rounds of feedback and user testing, I will continue to revise the existing dashboard and layout in future iterations of this project. As expressed previously, the Sunburst diagram has its strengths when it comes to visually representing the number of years of individuals’ residency in the Townhouse. Given the difficulty of creating this diagram in Tableau Public, however, it may be advisable to substitute another visualization such as that of the Icicle diagram (as recommended by the professor).
In addition to continuing to refine the filters and interaction, I want to continue to build out more robust links and connections to the original archival sources of the information. Whether it is linking to census records, the original leases, or what few older images we have of the Townhouses, it expands both the reach of our collections as well as connects them with another public platform. Given that there are approximately 174 residents between all the buildings over the 64 years, it will require a significant digitization effort to have a complete set of individual records linked to the map, however, smaller phases can be strategically implemented to increase the existing archival records connections.
Another future visualization route to explore is that of network mapping. Introducing data which defines their “role,” “department/company,” and “affiliation” opens up a new opportunity to generate networks maps of the various clusters and connections between certain occupations and roles among the residents. A preliminary network map with residents as the Source and departments or employers as the Target is reproduced below, which could reveal interesting trends and relationships with further grouping into categories and improved labeling and stylization.
