The digitization of cultural archives and historical records is opening up new avenues for biographical research and teaching. On the one hand, historical ouevres and lifework collections can be newly visualized for different audiences. On the other hand, biographical trajectories can be represented and analyzed in an unforeseen manner. But how do visualizations of life and work go together? With this paper we reflect on ways and means how to shed light on the life of artists or other historical actors by the means of metadata provided by their ouevre. By representing the works of Charles W. Cushman in the PolyCube framework for cultural collection visualization, we shed light on aspects of his biography in a geo-temporal and categorial-temporal information space. We discuss how these visual-analytical frames of reference could be combined for their mutual contextualization, and how they can be hybridized with textual sources to provide a multimodal, narrative framework of biographical knowledge exploration and communication.
Oftentimes, historical object collections containing
artifacts, assets or correspondences of artists, are bequeathed
to archives or institutions. These corpora can provide
researchers with numerous insights into an artist’s life and
work on various levels. However, such rich collections also
bring along thorough challenges for researchers in terms of
knowledge organization and information integration. They
frequently contain large amounts of objects and documents,
and the associated object information constitutes a complex
and multi-dimensional information space. Thrown into this
space, it thus is frequently difficult and time-consuming for
biography researchers to develop a macro-analytical
understanding. Also, as Sandweiss notes, ”the question is
whether the digitization of images and their new
accessibility over the Internet fundamentally alter[s] the
meaning of the original pictures or whether they instead give us
greater access to those original meanings. [...] We must
develop new interpretive skills to understand this new visual
medium”
In this paper, we explore two corresponding types of
”distant viewing”-techniques for cultural collections
The PolyCube project revolves around the question, how
multiple data dimensions (i.e. space, categories, and
relations over time) can be presented in an integrated visual
manner to support users in gaining a better understanding
of a given collection—a ”more integrated mental model”.1
Due to its relevance, we put specific emphasis on the visual
integration of time-oriented data. In the cultural heritage
domain time has been shown to be a crucial data dimension
Conventional visualizations of cultural collections
oftentimes build on ’coordinated multiple views’
Various visualization methods have been developed to
represent spatial and temporal data aspects in an integrated
fashion.2 In the following, we zoom in on geographic
space-time cube representations (in short geo-time cube),
which merge a map and a timeline in an orthogonal
fashion, to unfold a three-dimensional, cubic space
From a cognitive perspective, geographic space-time cubes
offer a specifically balanced design, by which temporal
and geographic origin are both mapped to a shared
display space, making temporal and spatial information
as2See the work of Krigl
In our implementation
In cultural collections, categorization is an important
technique used to organize and delineate the parallel existence
of genres, motifs, movements, or topics. Cultural
categories or taxonomies thus group artifacts into set-typed
ensembles—which also develop over time. The strengths
of space-time cubes in supporting the temporal exploration
of larger datasets can also be leveraged for non-geographic
data structures
To facilitate the interpretation of the temporal development
of individual sets, users can activate a ”hull” structure in our
implementation of a set-time cube, which connects the
geometrical vertices of each set over time. Missing data at one
point in time will show up as gap within this hull structure.
The shape of this hull allows users to easily trace the
temporal development according to the angle ( ) of sets and
shows basic flow patterns of set dynamics, such as
emergence, growth, diminution, and decline
3Position as a visual variable is known to be the easiest to
decode by users
Furthermore, users can interact with the visualization by
(de-)activating the hull, rotating the view, zooming and
panning, highlighting one or two sets, and selecting individual
objects for a preview and details
Similar to two-dimensional views, three-dimensional representations can also be combined to form ”coordinated space-time cubes”. As a coordinated assembly, multiple cubes offer a spatially proximate depiction of the temporal development of cultural collections in geographical and categorical space-time. To support users in linking these representations, interaction techniques (such as simultaneous temporal or categorical filtering or highlighting in both cubes) can help to explore parts of a collection in a coordinated fashion. In the following we will outline, how such a visual-analytical environment for object collections can also be used to shed light on the life of historical actors— and thus support the investigations of historians and biographers.
As a use case, we focus on the life and work of Charles Weever Cushman (1896-1972), who was a well-travelled amateur photographer and U.S. citizen.4 Visitors to the Cushman archive are introduced to his life with a short biography, recollecting essential facts on his background and activities.5 This text summarizes the rather sparse information known about his life, and points out the need for further enrichment by itself—a condition which applies for many historical records of historical actors.
Charles W. Cushman was born in 1896 in the
American Midwest (Poseyville, IN), studied law and worked for
different companies close to Chicago. He was married
twice—in 1924 and 1970—and died on June 8, 1972.
Stations of his winded life-path led him to work for a railroad
office (1917), in the Navy during World War I (1918,
honorably discharged from reserves in 1921), for Lasalle
Extension University (1922-1928), for a statistics office
(19281932), and for a brewery company (1932-1937). From
1937 until 1942 he was ”not employed [...] and had no
permanent residence. Traveled extensively”
5https://webapp1.dlib.indiana.edu/ cushman/overview/cushmanBio.jsp
6Due to the design of the utilized GeoTime package
It is not only for this blank period of Cushman’s life—but
also for an unknown and travel-intense segment between
1937 and 1942—that objects contained in his photography
collection can fill in rich information and shed light on his
biography7—and more general on life in the U.S. that
happen to see a massive economic, political and cultural
upswing in these times. Figure 3 models these two periods
of missing information as interruptions of Cushman’s
biographic pathway. While we also see a need to honestly
represent uncertain or missing data in many areas in such
an explicit fashion
The Indiana University—to which Cushman bequeathed
his lifework—features a digitized collection of 14.500 of
Cushman’s Kodachrome photographs from 1938 to 1969
7As Martha Sandweiss (2007b) formulates it: ”Historians
more often confront the difficulties of interpreting images
without extensive biographical information on the photographer than
those of interpreting pictures in light of the photographers own
readings of them.” (p. 194)
Cushman consistently documented his activities by the
means of a travelogue, listing locations and dates for all
his photographs together with a short content description.
This information has been transcribed for all pictures, and
the resulting catalogue has been further enriched with
geocoordinate
To gain a better understanding of Cushman’s photography collection, we use the PolyCube system to visualize the geographical and categorial patterns of the photographs with their temporal developments between 1938 and 1955. We will first discuss the geo-temporal dispersion of the photographs as indicator of Cushman’s travel activities, and emphasize new information which they provide to fill a good part of both blank periods in figure 3. Then we will take a closer look at the categorial-temporal developments within the collection, before we finally demonstrate how even higher-dimensional questions can be explored by means of coordinated multiple cubes.
Figure 4 introduces a geographic space-time cube
representation of Cushman’s photographs between 1938 and 1955.
In this representation, early photographs are shown at the
bottom (violet) and the latest ones at the top (yellow). The
widely scattered positions of the pictures near the bottom
document Cushman’s extensive travels through the United
States between 1938 and 1942. For this period of time,
the online biography just recites Cushman’s own statement
on his application form for Federal Employment: ”Not
employed during this period and had no permanent
residence. Traveled extensively. Looked after personal
interests, principal of which was interest in contract of sale of
Drewry’s business”
From 1943 onward, Cushman remains quite stationary
within the Chicago region—even after his federal
employment in the War Department of the Chicago ordnance
district ended. An obvious explanation could be the
shooting and his wife’s following hospitalization at a sanitarium
in Oconomowoc, WI
Figure 5 depicts a categorial set-time cube representation
of the same corpus selection based on the primary genres
of the photographs. The visualization shows that across all
photographic categories, Cushman took the fewest pictures
in the year 1943. Their number increased again over the
following years, with an especially high number in 1952—
the year the couple moved over to the west coast.
With an activated hull structure, the set-time cube makes
the main subject of Cushman’s pictures visible as they
develop over time. While he took a wide variety of
photographs before the shooting, the gaps in 1943 indicate that
he took only ”snapshots” (dark green in figure 5,
”landscape” (red) and ”identification photographs” (light orange,
which also mainly contained plants and buildings) and
did not regain his full variety until 1946. A category of
photographs he hardly used after 1942 are ”glamour
photographs” (in light blue on the right in figure 5).
Similar to 1943, in the year 1948, a gap within many
genres can be observed. Studying the archive’s biography
The PolyCube framework also offers a parallel setup of
”coordinated multiple cubes”, which allows to use the
photographic genres as filters for the geo- and the set-time
cubes to explore more specific questions. To illustrate this
option, we took a closer look at landscape photographs that
show an initial decline in numbers in 1942 and increase
massively again in 1952 (Figure 6, right). Selecting this
category we can zoom and filter on this subset of
landscape pictures, and explore their spatio-temporal
distribution (Figure 6, left). A large number of landscape
photographs was taken at the West Coast especially in 1954
and 1952 as well as in the Middle West in those years.
The amount of landscape photographs during Cushmans
earlier years is linked to his travel activities as well.
Doing the same analysis with other categories, we can deduce
that Cushman took the largest number of architectural and
cityscape photographs during his time in the Chicago area.
Though this might not be surprising and evokes the image
of a contemporary tourist, this analysis points to a simple
fact: As Cushman was an amateur photographer the
subjects he chose were not dependent on any professional
focus or on his clients demands, but on his own interests. ”He
crafted an extraordinarily complete inner world from the
fragments of life that he found around him, a world defined
more in terms of his own experiences than in the service of
a search for a hidden essence or an unknowable design”
With this paper, we explored the potential of a visualization method for time-oriented collection data to enrich the historians and biographers methods portfolio. Deliberately going beyond one-dimensional data portraits, it provides the means for an integrated ”distant and close reading” inspection of large cultural collections from two time-oriented macroanalytical perspectives.
In a case study on the American amateur photographer Charles W. Cushman, we showed both how object data can fill in periods of lacking biographical knowledge, and more specifically, how the development of a collection could be inspected for consequences of critical life events—on a geographic and on a categorial level. We also discussed how the pictures show no direct connection to Cushman’s professional life and rather appear as a resource for further exploration and understanding of his private life.
As with many case studies in the digital humanities realm, we consider the results to indicate a possible general application area, while remaining a local exploration until further notice. Obviously, due to the idiosyncratic data structures in historical fields of study, the analytical avenues generated in this case study cannot be directly transferred to other archives or other people’s lifework collections. However, due to the wide-spread use of temporal, geographic, and categorial descriptors, we consider the PolyCube framework to provide a relevant macroanalytical extension for biographical research. Distant reading or viewing approaches like the ones discussed in this paper offer dynamic overviews on large archival collections. These representations again enable users (1) to identify and interpret basic patterns within a lifework collection on an macroanalytical level and (2) to identify objects of interest for more detailed investigation, e.g. with traditional arthistorical, formal, or also critical-interpretative means.
Visualizations obviously depend on digitized content and can represent only those events of a biography, where corresponding data or documents exist already in a digital format. Thus, for the purpose of extended biographical research, our inspected data selection (from 1938 to 1955) would require a substantial extension—for example, with information on events which happened before and after our archival sample. Also information on critical life events might have been omitted from the official biography on purpose—and thus would require additional investigation and corroboration.
We consider the added value of visual analytical
approaches to historical records to generally hinge on a
minimum level of data quality and data specifity. Only very few
historical object or lifework collections lend themselves as
a completive source for the creation of high resolution
portraits. For this purpose, only specific artifact collections
provide either detailed or consistent enough documentation
in multiple metadata dimensions (such as on time of origin,
place of origin, category) for each object. From a
generalization point of view, this is no negligible restriction, as
only modern technologies generate these inscriptions
reliably by themselves (e.g., digital photography)—and only
few artists pay attention to meticulously document the
circumstantial or contextual conditions of their creations.
In all other cases, uncertainty and incompleteness of
collection data is a major challenge—and visualization has to
find new ways to cope with these standard conditions of
historiographic data—and the humanities’ knowledge
consistency in general
Going one step further, biography visualizations (depicting a person’s direct spatio-temporal trajectory like in figure 3) and lifework visualizations (depicting objects like in figure 4 and 5) could be beautifully combined to shed light on each other. Future work on the PolyCube framework will aim for the simultaneous and consistent visualization of both types of complex dynamics. In our analysis of the Cushman collection, we found it very useful for our interpretation of his lifework to relate the photographs to events in Cushman’s life. But at other points, we figured that the lifework collection can provide us with information going beyond the biographical knowledge—e.g., giving more detailed information on his travel activities or changes in Cushman’s photographic interests, which could hint at another (yet unknown) critical event in his life.
We consider such contextually rich and transactional
approaches to interpretation to be of relevance in multiple
history-oriented knowledge domains. Even ongoing
debates with proponents of a ”new criticism”-style focus on
objects or ”texts themselves” seem to benefit from the
emergence of these novel methodological options, as they
help both micro- and macroanalytical sides to redefine and
recalibrate their positions—including the development of
research programs which systematically aim to intertwine
and mediate the best of both interpretive worlds
This research was supported by a grant from the Austrian Science Fund (FWF), project number P28363-G24, from the NFB, project number SC16-032, and the European Union H2020 research and innovation programme under grant agreement No. 101004825.