Blog Post

Spring Spotlight #4 - Digital History Project Management at Stanford

Spring Spotlight #4 - Digital History Project Management at Stanford

DIGITAL HISTORY PROJECT MANAGEMENT

@ STANFORD

 
Spotlight organizer: Katherine McDonough


Featured Practitioner: Zephyr Frank, Associate Professor of Latin American History, Stanford
Project: Spatial History Project, Stanford University

Featured Practitioner: Elijah Meeks, Digital Humanities Specialist, Stanford
Project: Orbis: Stanford Geospatial Network Model of the Roman World

*NEW* Featured Practitioner(s): Paula Findlen/Dan Edelstein/Nicole Coleman
Project: Mapping the Republic of Letters, Stanford University

 

 

ZEPHYR FRANK

SPATIAL HISTORY PROJECT | TERRAIN OF HISTORY-RIO DE JANEIRO HISTORICAL GIS

Question 1:Project description: Can you describe the origins, support for, and present status of the Spatial History Project? From data exploration to project publishing, how have you and your team used digital tools to pursue this project?

The Spatial History Project traces its origins to a Mellon Distinguished Scholar Fellowship awarded to Richard White, my History Department colleague, in 2007.  The award was used to found the project and provide funding for professional staff and research support for the first three years of the lab’s existence.  Owing to careful use of the funds, our lab was able to extend the grant money into a fourth year (2010-11).  Meanwhile, the SHP was able to find space on the 4th floor of Wallenberg Hall, in the central quad of the Stanford campus, in a research space controlled by the Dean of Research.  This location proved quite important to sustaining the SHP as it connected our lab to the Dean of Research and the activities and funding associated with Wallenberg Hall via the Wallenberg Foundation.  A major gift from the Wallenberg Foundation in 2011 allowed us to initiate a program of postdoctoral researchers and to build international ties to universities in Sweden.  In particular, we have strong ties with the Humlab at Umea University, which has provided us with advice and inspiration as we’ve developed our lab over the past several years.  Finally, our lab has been awarded a second Mellon Foundation grant for the study of crowdsourcing in humanities research.  This last grant was awarded under the auspices of CESTA and includes the participation of the Literary Lab as well as SHP.

As for digital tools, this is where my own research really comes into the picture.  Starting back in 2003, I began to learn how to use GIS software as part of my research project on 19th-century Rio de Janeiro.  Building on this interest, I started adding spatial history to my teaching repertoire and, in this way, became one of the faculty colleagues Richard White could turn to when he decided to found the SHP.  Along with GIS, which Richard was using for his railroads project and I was using for Rio, we also began to use Adobe Flash for the creation of interactive visualizations.  The GIS software (ESRI’s ArcMap suite) tended to provide the background spatial data, which we then processed and cleaned up in Illustrator/Flash for publication on our website in the form of visualizations.  In recent years we have also begun to use network analysis/viz software—Gephi in particular—as an additional tool for visualizing and exploring social space (e.g. networks of boards of directors of railroad companies, though this was not with Gephi).  Last, I would mention our use of Tableau, a data viz and analysis program developed here at Stanford by colleagues such as Jeff Heer in CS.  Tableau has been a wonderful digital tool for the exploration of large datasets, providing a useful supplement to our GIS tools and letting us view our data quickly and in multiple formats.

Question 2: People/Space: How have different on- and off-campus collaborators contributed to the project? What role have "lab" environments played in the pursuit of this project?

On-campus collaborators have played a significant role.  These extend from the many student research assistants, whom we’d certainly characterize as collaborators and team members, to faculty who’ve participated by developing their own projects.  Our projects are inherently collaborative, team-based, and reliant on the combination of professional staff and research space afforded by our lab.  Faculty contribute by bringing their research questions to the lab.  Students contribute by bringing their insights and creativity to bear upon these questions.  Staff (the list is long, primary staff have included Erik Steiner, Maria Santos, Kathy Harris, Mithu Datta, Jake Coolidge, Whitney Berry, Killeen Hansen, Celena Allen, and Jeremy Zahlen) provide the glue that holds our teams together; they help take ideas from the whiteboard to polished online visualizations.  Off-campus collaborators have done much the same things.  We have hosted faculty from Berkeley (Scott Saul, Linda Lewin) to Brazil (Sidney Chalhoub, Iris Kantor, Douglas Libby), Maryland (Daryle Williams) to Iowa (Lea Vandervelde), British Columbia (Henry Yu) to Colorado (Greg Simon), as well as our close collaborators from Umea University in Sweden (Patrik Svensson, Thomas Nygren, Fredric Palm).  We have also hosted undergraduates and graduate students from institutions such as Berkeley, Princeton, Harvard, Middlebury, the Federal University of Minas Gerais, Brazil, and the University of Colorado.  The outside faculty and students contribute projects and expertise—they extend our range of interests and connect us to DH research being done around the country and the world.  In all of this, it is safe to say that the lab environment has been critical to our success.  It provides the space and the people (staff and students in particular) without which none of this would really be possible.  The lab space offers opportunities for cross-project collaboration (including with affiliated labs under the CESTA banner).  

Question 3: Evidence: What are some of the challenges you have encountered when working with incomplete/uncertain data from archival sources. In what ways have you developed solutions deal with gaps in historical data?

There are really two levels to this question.  To paraphrase Don Rumsfeld, we have known unknowns and unknown unknowns.  There’s incomplete and then there’s really incomplete.  For instance, is a City Directory incomplete because it does not list every business?  Because it does not list many women?  Because it lists fewer kinds of business in 1849 than in 1889?  Or because half of it is simply lost?  I don’t think any of these issues is particular to digital methods, though the effect of incomplete or uncertain data can be of particular importance when we are talking about an interactive map or a database.  

Question 4: Goals: What were the initial goals and how have your research questions changed over the life of the SHP?

The initial goals of the SHP were to support the use of digital tools, such as GIS, in historical research, particularly on railroads in the US and the city of Rio de Janeiro.  Over time, the goals expanded to include other research projects and to the creation of ties with allied labs working in DH.  This expansion has certainly led to changes in the nature of my own project. From a largely quantitative and empirical study of Rio de Janeiro, focusing on maps, my project has moved toward the analysis of social spaces through the study of literary sources, in step with my increased collaboration with Franco Moretti and his associates in the Literary Lab (part of CESTA).

Question 5: Impact: Has your project had any impact in your field so far (publications, conferences, new collaborations, development of similar projects?), and how do you envision its long term contribution to the discipline of History and the Humanities more broadly?

The SHP has already had an impact on the field(s) of history and DH.  Richard White published his book Railroaded (Norton, 2011), which drew very heavily on work done in the lab.  In my own research, I published an article on the space of buying and selling slaves in Rio de Janeiro with an online visualization component in the Journal of Latin American Geography (2010).  But the biggest impact, I would say, has been with the many projects that have sprung up since the SHP’s inception on 2007.  That is, beyond Richard’s and my projects.  In this regard, the impact can be measured in terms of scores of interactive visualizations, thousands of hours of collaborative work, and the development of a couple-dozen research projects under the aegis of our lab. These projects are the true measure of our impact: they have helped transform the work of our colleagues and collaborators and have provided examples and inspiration for scholars well beyond Stanford.

Question 6: Future: How do you see Digital History evolving at Stanford, and how might graduate students and new faculty prepare for and participate in these transformations? 

DH is going through a period of growth at Stanford.  This entails the development of new initiatives, such as the new DH center envisioned in Stanford Libraries, as well as the continued bubbling up of individual, faculty or grad student initiated projects.  As for preparation, this really depends on the combination of technical chops (programming, stats, etc.) and background/domain expertise required by individual projects.  Participation is all about seeking out opportunities to collaborate.

 

 

ELIJAH MEEKS

ORBIS

Question 1:Project description: Can you describe the origins, support for, and present status of Orbis? From data exploration to project publishing, how have you and your team used digital tools to pursue this project?

ORBIS began as an attempt to create a cartogram similar to Tom Carden's map of the London Subway, except for the Roman Empire. Walter Scheidel, the PI for ORBIS, wanted to demonstrate that the distance of place from place in the Roman world had little to do with its geographic distance and more to do with its network distance. In order to accomplish this goal, we needed to develop a dataset of sites--most of which came from Pleiades, routes that connect them--which we drew from the Barrington Atlas in the case of roads, modern river tracks, and modeled sea routes. Stanford Library supported the project by attaching my position to Walter's research for a nine month period. As we entered the final stage of the project, it became clear that the results needed a dynamic, interactive form of publication, and so with Karl Grossner's assistance, we built the ORBIS site. I'd like to stress here that data exploration was not nearly as important as systems or model exploration. Much of the value of the project came not from looking at the data, which except for the sea model was already well-known, but from theorizing the formal connections between the data and developing a model or system to represent those connections.

Question 2: People/Space: How have different on- and off-campus collaborators contributed to the project? If applicable, could you evaluate the role of "lab" environments in the pursuit of this project?

There was no lab involved with ORBIS. Off-campus collaboration came in the form of direct communication between the PI and other PIs doing similar work in the field, or between myself and other scholars or developers who were working on similar projects. Since the publication of ORBIS, there have been significantly increased opportunities for collaboration, not only in Classics but in environmental science, pedagogy, information visualization, and the other spheres that ORBIS touches on.


Question 3: Evidence: What are some of the challenges you have encountered when working with incomplete/uncertain data from archival sources. In what ways have you developed solutions deal with gaps in historical data?

One of the key concepts of a model like ORBIS is that you purposefully simplify highly complex systems, like the transportation system of Rome. As a result, the product of ORBIS is a generalized view of movement of people and goods in the period and region that it covers. The scale, particularly, affects results, with shorter routes having higher variation from historical records than longer routes. But ORBIS was meant to provide a strategic and aggregated view of Rome, and at that scale performs well. Moving forward, we plan to integrate the historical and systemic uncertainties and variation, that are known and discussed in the text, into the actual interface and representation. This involves working with probabilistic methods and poses challenges not only in analysis but representation.


Question 4: Goals: What were the initial goals and how have research questions changed over the life of Orbis?

The initial goals were to provide an overall view of the distance of the rest of the Roman World from a particular point within it. The unintended consequence was that we could easily develop a route-finding interface that proved extremely accessible. We also discovered that by leveraging PostGIS and pgRouting we could explore much more sophisticated questions of cost to travel. The requirement to represent sea travel necessitated a large-scale engagement with modeling wind-driven sea transport, and resulted in an extremely valuable contribution to understanding travel overall, as well as revealing significant areas for improvement of historical state-scale sea travel.

Question 5: Impact: Has your project had any impact in your field so far (publications, conferences, new collaborations, development of similar projects?), and how do you envision its long term contribution to the discipline of History and the Humanities more broadly?

ORBIS was featured in several major magazines and news papers and received a great deal of attention by the Internet community at large. It has also resulted in several papers, many talks, and a mini-conference where scholars in the Classics and elsewhere discussed the application of such methods to other geographic regions as well as different periods, along with discussing improvements to the basic model. ORBIS and/or its data has been used by several different collaborators for historical climate research, pedagogy, and traditional scholarship in the Classics.


Question 6: Future: How do you see Digital History evolving at Stanford, and how might graduate students, alt-ac staff, and new faculty prepare for and participate in these transformations?

I think the digital tools and methods for doing Digital History have reached a critical point of being highly accessible and comprehensible for non-experts. I also think that the resources for doing Digital History at Stanford are particularly extensive, with the Digital Humanities Specialist, Academic Technology Specialist for History, Spatial History Lab, and Areas Studies Librarians all deeply conversant with cutting edge digital history practices. These factors, coupled with the increased visibility of digital history publications, should prove extremely attractive to faculty, graduate students, and allied alt-ac positions. However, it can be daunting for a new faculty member or graduate student to consider the investment in doing digital history given the time it might seem to take away from traditional scholarship. With that in mind, I would suggest such scholars look to elements of their ongoing research for which they have deep interest in, and interface with support staff to find out if there are existing data to improve that research using computational methods and/or digital publication, and only then pursue such research. Critically, I think scholars considering such a move need to keep their focus on describing historical processes and systems, and not create datasets. It is in the formal description of relationships between pieces of data where research and scholarship is most vibrant, and not in the collection and annotation of those individual pieces.

 

 

 

NICOLE COLEMAN, PAULA FINDLEN, DAN EDELSTEIN

MAPPING THE REPUBLIC OF LETTERS

 

Question 1:Project description: Can you describe the origins, support for, and present status of Mapping the Republic of Letters? From data exploration to project publishing, how have you and your team used digital tools to pursue this project?

This project grew out of a fairly tradiitonal academic conference on the Republic of Letters, held in 2007. At the end of the conference, we realized that each presenter had talked about a small corner about the Republic of Letters, and that we still had no overall picture of the whole. A group of Stanford faculty got together to think about this problem further. At first, we had no plans to take the project in a DH direction. But we got a lucky break with the Robert McNamee, at the Electronic Enlightenment Project at Oxford, agreed to give us the metadata for the correspondences they'd digitized. We were then able to devise a visualization that let us see large patterns in correspondences. Since then, we have also acquired many other datasets, including the metadata for the correspondences of Kircher, Franklin, d'Alembert, and others.

We currently use biographical data, correspondence data, and travel data to track the movements of people and ideas, primarily in the 17th and 18th centuries. From the early stages of the project we decided to build tools that would run natively in a web browser so that we could easily disseminate both the tools and the research via the Web. Visualizing large data sets has given us the opportunity to see people, letters, and movement in a much larger context that was possible through reading the letters.

The project received a three-year grant from the Stanford Presidential Fund for Innovation in the Humanities and soon after was awarded an NEH Digging Into Data grant. Throughout, the Stanford Humanities Center has supported our work and we now receive funding from the Dean of Research, the Vice Provost for Graduate Education, and the Vice Provost for Undergraduate Education.

Question 2: People/Space: How have different on- and off-campus collaborators contributed to the project? Could you evaluate the role of "lab" environments in the pursuit of this project?

Our lab, originally at the Stanford Humanities Center and now at the Center for Spatial and Textual Analysis, has been the hub of activity for the project and essential to the project's success. We meet regularly in small working teams and with the full group to review work in progress. It's also the place we bring visitors for presentations of our work.  The lab environment is essential for this type of collaborative work which requires frequent ad hoc meetings and review sessions. We have had a number of off-campus visitors work with us for months at a time in the lab. It is also the research home to our postdoctoral researchers.

We have built long-standing relationships with partners doing complementary work in correspondence: "Cultures of Knowledge" at Oxford and "Circulation of Knowledge in 17th c Dutch Correspondence Networks" in the Netherlands. We have also established a close working relationship with the DensityDesign Research Lab at the Politecnico in Milan.

Question 3: Evidence: What are some of the challenges you have encountered when working with incomplete/uncertain data from archival sources. In what ways have you developed solutions deal with gaps in historical data?

Visualizations of data play an essential role in data exploration and analysis by enabling us to plot many data points on a single map, graph or timeline. Many of the existing tools, though, do not provide the kinds of functionality that humanists require, in particular the ability to perform more qualitative (i.e., interpretive), rather than purely quantitative (e.g. statistical) analysis. One reason why this qualitative approach is essential is that our data vary considerably from most scientific data. Historical data are generally incomplete and uncertain, and what’s more, in unpredictable ways. Our response to this has been to develop tools that show the gaps in the data and tools that leave room for domain experts to interact and manipulate the data directly through the visualization. In the first case, if we have dates for letters but are missing locations we will plot those letters on a timeline that corresponds to the map showing letters for which we have coordinate locations. We can extend this to show the language of the letters, the gender of the correspondents, nationality, and about any other biographical dimension of the correspondents. In the latter case, we make it possible for domain experts to define their own categories to apply to the Republicans. Where "nationality" and "profession" are not applicable, we have developed categories like "national-cultural identity" and "interests." 

Question 4: Goals: What were the initial goals and how have research questions changed over the life of Mapping the Republic of Letters? 

Our broad goal was to gain a better understanding of the Republic of Letters through individual case studies. We began exploring the spatial dimensions, but soon realized that do to the ambiguity and incompleteness of the data, geo-spatial mapping was both very limited and at time distorted our understanding of the exchange of ideas. We gradually began to think more and more in terms of connections between people defined by place, affiliation, and more explicit relationship, such as correspondence. A more precise understanding of how these networks were constructed will allow scholars to tackle questions that have long hovered over the field of intellectual history, but remained tantalizingly out of reach. Were intellectual networks truly cosmopolitan, or did they tend to be nationalist in breadth? What role did social elites and civil servants play in the constitution and diffusion of intellectual networks? Did these networks look considerably different in the sixteenth versus the eighteenth century? How do intellectual networks map onto other kinds of networks (trade, diplomatic, colonial, etc.)?

Question 5: Impact: Has your project had any impact in your field so far (publications, conferences, new collaborations, development of similar projects?), and how do you envision its long term contribution to the discipline of History and the Humanities more broadly?

Our project has received a large amount of publicity, both in scholarly journals and in the press (it was featured in a New York Times article about "Humanities 2.0"). We're currently writing up a series of articles for publication in a major history journal. 

Question 6: Future: How do you see Digital History evolving at Stanford, and how might graduate students, alt-ac staff, and new faculty prepare for and participate in these transformations? 

We feel strongly that all graduate students in the humanities should come out of Stanford with a basic literary in digital technologies, whether or not they currently use them for their research. While we can imagine many different ways in which DH will be integrated into the curriculum and research, our own hope is for these approaches to become a standard part of every scholar's toolkit -- to the extent that in 10-15 years, we no longer need to distinguish between "digital" history and history per se.

 

93

1 comment

As someone who is just getting her feet wet in spatial history, I find both of these projects incredibly interesting and instructive. A few ideas came out in both that I'd like to highlight:

1. The critical importance of collaboration. Both in developing technical competency for working on such projects as these, or in knowing who to ask for help, collaboration is vitally important in doing large-scale projects. 

2. The limitation of simplification. I think we all know that a representation is not the thing itself (the map is not the territory). But in beautiful and complex projects such as the Spatial History Project or ORBIS, it's easy to think (from the outside) that all the variables have been accounted for, all the nuances have been successfully depicted. And I think what we see from both Zephyr and Elijah is that mapping can't account for everything. But it's really important to know what the representation does and does not depict: Are you modeling potential travel routes instead of "actual" ones? Does your city directory include all the businesses? I like Zephyr's idea of the "known unknowns and the unknown unknowns." If you know that you've left things out (and you have, always), then you can account for them. In DH, especially, accounting for the unknowns is vital. 

 

 

108