German Networks on the Third Crusade


As we are increasingly confronted with actors operating outside “normal” state-centric networks, studying a period when states paradoxically consisted of non-state networks can be very enlightening. I have mentioned the study of medieval England in a previous post. My third historical network analysis is that of the medieval German Empire, in particular during the crusades.

One of the largest military undertakings of the Middle Ages was mobilization of the Third Crusade, the prelude to which was the subject of the film Kingdom of Heaven 2005. In July of 1187, Saladin, ruler of Syria and Egypt and founder of the Ayyubid dynasty, had destroyed the armies of the Christian kingdom of Jerusalem at the Battle of Hattin. When confirmation of this defeat reached the western European powers, they began preparations for a campaign to aid the kingdom, and continued with greater urgency when confirmation arrived of the loss of Jerusalem. What audiences today remember most about the Third Crusade is the military contest between Saladin and Richard “the Lionheart” of England, usually as the backdrop to Robin Hood fiction.

However, armies from the German empire played a significant, if ultimately truncated, part, in the counter-offensives that were launched in 1189 and 1190. The main imperial army, under Emperor Frederick I, is thought to have numbered about 15,000 fighting men, and marched overland through the Balkans and modern-day Turkey, surviving multiple enemies until the emperor accidentally drowned within a couple weeks of Syria (it ranks as one of the most ironic deaths of the Middle Ages). Other groups sailed from North Germany, or from Italy, and arrived in time to join the fighting around Acre, in modern-day Israel.

German scholars have long recognized that networks determined the nature of power and its application in the medieval German empire–we have some brilliant monographs on the elite who had dealings with the German emperors, and what the patterns of these associations mean. [1] However, to date no one has shown much interest in using network science to analyze this data, although several long-standing debates revolve around the very questions network science is designed to answer. For example, in the German realm especially, political power was contested and particularly influenced by faction and interest. It follows, then, that any German monarch leading a large army out of Germany would use troops drawn from his own lands particularly, and not so much troops from his rivals or those dissatisfied with the power or behavior of the emperor’s family.

Continue reading

Posted in Research Projects | Tagged , , , , , | Leave a comment

Newsblast Volume 5 Issue 1


In the current issue of the Network Science Center Newsblast Dr. Daniel Franke discusses a project he is working on, sifting through personal testimony and official reports to create command networks during the Civil War. He is studying the Overland Campaign in 1864 and he will subsequently be exploring the command networks of the Vicksburg Campaign of 1863.  To learn more read the article, ‘Networks of Command: The Battle of the Wilderness, 5 May 1864′, in this issue of the Network Science Center’s Newsblast click here.


Posted in News | Tagged , , | Leave a comment

Networks and Command in the American Civil War


How might Network Science impact the study of the Civil War? One way, which I have been exploring this fall semester, is by analyzing the transmission and reception of information by commanders at key moments in the military events of the war. In particular, the Overland Campaign of May 5 to June 7, 1864, could yield interesting facts once the parameters of information flow have been categorized and organized in a way interpretable by analysis programs such as Gephi and ORA.

“Overland” was perceived at the time as a titanic, horrific event in American history, and it still engenders controversy today. Ulysses S. Grant, commanding the armies of the United States, launched his c. 120,000-strong armies against the c. 60,000-strong army of Robert E. Lee with the aim of destroying the Confederate army as an effective fighting force, or at the very least taking the initiative away from the intimidating rebel general. The opening rounds of the ensuing contest took place in the vicinity of the old Chancellorsville battlefields, at the Wilderness on May 5-6, and at Spotsylvania Court House on May 8-21. In both instances, the Union Army came very close to breaking the Confederates completely, but in the end Lee’s army held while inflicting heavy casualties on its Federal enemy. By the end of the campaign at Cold Harbor in early June, Confederate casualties had surpassed 30,000, and Union casualties were upwards of a shocking 55,000.

Currently I am working on three major historical questions that can be addressed through network analysis. One, not surprisingly, is “who were the central nodes” in the Wilderness and Spotsylvania? By this I mean both who had the greatest amount of information pass through their hands, and who made the key decisions in the battles. Second, what was the commanders’ state of knowledge at any given time during these battles? And third, how did the leadership of the Union’s Army of the Potomac change during the campaign, from the regimental leadership up?  A very pertinent fourth question would be how our understanding of the campaign is altered as a result of network analysis.

Establishing the parameters of the data set is fairly challenging, both because the data by its very nature is incomplete and because the data that is chosen depends largely on one’s assumptions going into the project–historical data is not always clearly defined. It also underscores the discussions and debates among historians about how Civil War generals actually received and transmitted information. [1] By way of brief example, if we examine the opening of the Battle of the Wilderness on May 5, 1864, we are confronted with the immediate question of whether every message from one headquarters to another should count as an edge connecting headquarters’ nodes (which, depending on the question asked, could be a valid approach), or whether every person who handles the message should count as a node. The latter is both more precise and, as Civil War historians would agree, more reflective of the fact that the commander himself could well have been absent from headquarters when crucial messages arrived. It also allows us to notice anomalies, such as the famous instance when George Meade, commanding the Army of the Potomac, bypassed the chain of command, including his own chief of staff, to issue orders directly to a division commander. To complete this kind of network, however, requires a close mining of memoirs and letters as well as official reports, not to mention some deductive reasoning, such as determining if a meeting of commanders had to have taken place even though there is no actual mention of it anywhere. And then there is always the question of how to graph Grant’s relationship with Meade, summarized so succinctly by Meade’s staff officer Theodore Lyman in his letters “General Grant had his station with us (or we with him).” [2]

The issue of how high casualties affected unit cohesion is a somewhat different and larger issue, and one that will occupy much of my spring term of 2015. In the mean time, we keep moving “by the left flank.”

[1] Significant studies of command and control in the Civil War include R. Steven Jones’ The Right Hand of Command: Use and Disuse of Personal Staffs in the Civil War (2000); Richard L. DiNardo’s “Longstreet and Jackson Compared: Corps Staff and the Exercise of Command in the Army of Northern Virginia,” in Longstreet: The Man, The Soldier, the Controversy, ed. R. L. DiNardo and A. A. Nofi (1998); and Archer Jones’ Civil War Command and Strategy: The Process of Victory and Defeat (1992).

[2] COL Theodore Lyman, Meade’s Headquarters 1863-1865, Boston 1922, page 91.

Posted in Research Projects, Uncategorized | Tagged , , , , , , , , | Leave a comment

Tech Hub Sustainability in Africa #HubSustainability

Researcher Dan Evans continued to participate in a series of Google Hangouts partnering with AfriLabs and AfriHive discussing tech hub sustainability in Sub-Saharan Africa. The most recent occurred on 5 December.

The main focus of these discussions have been the following question: “Is it possible for tech hubs to strike a balance of achieving their mission and ensure sustainability?”

I’m probably oversimplifying a bit, but a common story among discussants is:

  1. There was a great deal of optimism when the hubs were established.
  2. The hubs developed business/growth plans that aggressively assumed that after the initial funding from a typical traditional Development Finance Institution (DFI), the hub would soon be incubator/accelerator-like and the revenue from start-ups working out of the hub would quickly allow the hub to achieve sustainability.

Now, we are a few years down the road and reality is hitting. Based on the local ecosystems, some of the hubs have executed the dreaded “pivot.” In some cities, the community really needed an organization that allowed tech enthusiasts to congregate and interact, as well as receive tech training-an important mission, but not an activity that leads to near-term sustainability. Simply, the expected start-up ecosystem has not emerged-yet.

It seems from speaking with hub management teams, that the original funding organizations are now looking for expected returns but they are not forecasted until sometime in the future. We recently saw this situation during our August visit to the Mara hub in Dar es Salaam when the sponsoring funding agency radically modified the business model.

As part of my ongoing research project, I am attempting to ascertain if there are business models that are more effective than others. For example, the Meltwater Entrepreneurial School of Technology (MEST) in Accra, Ghana is well-funded by the Meltwater Foundation, while hubs in Tanzania and Rwanda have official government backing, and other hubs across the continent are founded by traditional DFIs. Because we are still early in “the story,” we don’t know the answer-time will tell.

My next blog post will discuss some ideas that might assist in finding this balance for future “hub and funder” partnerships.

Posted in Research Projects, Uncategorized | Leave a comment

Newsblast Volume 4 Issue 12


In the current issue of the Network Science Center Newsblast Dr. Jocelyn Bell discusses an interdisciplinary research project she is working on involving pulling meaning from machine read text using significant words and N-grams.  To learn more read the article, ‘Graphs, Maps, and Palm Trees: Constructing an Associative Literary Network’, in this issue of the Network Science Center’s Newsblast click here.


Posted in News | Tagged | Leave a comment

The Best Seat at the Table


Repost of study submitted two years ago by then Cadet now 2LT Jeffrey Nielsen

With Thanksgiving fast approaching some people might be wondering, what is the best seat at the Thanksgiving table? Two years ago as an independent project, CDT Jeffrey Nielsen used a network science approach to find the best seat at the mess hall table here at the United States Military Academy. Cadets are served their meals family style at tables with 10 seats and have only a limited amount of time to finish their meal. Condiments, fixings, sides, and drinks are all placed on the table to be shared. There are also duties that cadets in certain seats are responsible for. His findings might help you decide where to sit at your Thanksgiving dinner table.

The family-style tables of the cadet mess hall feature 10 seats and 4 distinct food clusters: condiments, fixings, sides, and drinks. During each meal, cadets will have to frequently request and then pass food items to each other. If each cadet is an agent node and each food cluster is a resource node, then the best seat is the one with the highest degree centrality (so that you can easily receive food) and the lowest betweenness centrality (so that you don’t have to pass any food).

In order to construct the table network, Cadet Nielsen created a diagram of each cadet’s arm length a few feet over the table. If a food cluster was within a cadet’s arm length, then there was a link between them, and likewise for any two cadets whose arm lengths overlapped. For example, the table commandant (seat 1) can reach the condiments as well as seats 2a, 2b, 3a, and 3b because they’re within arm’s reach. The final addition to the network was duties. The plebes (first year cadets in seats 5a, 5b, and 6) at each table are required to perform certain table duties as well as recite knowledge before the table commandant allows the table to begin eating.

After building the table network in ORA, the centrality measures of degree and betweenness were calculated. The seats were ranked in descending order of degree and ascending order of betweenness to reflect the “best seat” assumptions and these two values were then added to reach a final ranking. The best seat is a tie between 3a and 3b: an unsurprising conclusion considering network symmetry, but a very surprising conclusion in that the Cows (third year cadets) have the best seats at the table! Click here to see his full presentation.

This year when you sit down to your Thanksgiving feast consider Cadet Nielsen’s results and you can choose the best seat at the table.

Posted in Research Projects | Tagged , | Leave a comment

2014 Cadet Seminar on Network Science


On 15 November a group of 30 high school students, parents, and teachers from local schools met with cadets and faculty from West Point to learn more about network science. Lori Sheetz welcomed the group with an overview of the Network Science Center and some of the research it conducts at West Point, then Cadets Will O’Keefe, Lewis Black, Ioannis Wallingford, Molly Prins, and Jake Moffatt gave brief talks explaining their projects. The purpose of the seminar was to give the high school students a chance to talk network science with near peers and to learn about more applications of network science accessible to high school students. The parents were invited to gain a better understanding of the type of research their children are working on and other teachers to generate interest for the future projects. After the talks Cadet Lewis Black, President of Social Network Analysis Club (SNAC), shared with the students some of SNACs current activities and how the cadets have organized the club. SNAC represents the transition from taking a formal network science class, to using the same tools in independent research, and finally to an informal gathering of students engaged in further exploring applications of network science. In lieu of a formal class the high school teams participated in the summer workshop in Boston and are currently working on a yearlong research project. The next objective would be to see if the teams translate that experience into tools that they use independently to solve real world interdisciplinary problems.

Cadets Ioannis Wallingford, Molly Prins, Jake Moffatt, and Antoine Davis took the visitors on a tour of central post area and explained the cadet honor code, trophy point and daily life of a cadet at the academy.

Posted in News | Tagged , | Leave a comment

Tech Hub Sustainability In Africa #HubSustainability

Researcher Dan Evans will participate in a series of Google Hangouts partnering with AfriLabs and AfriHive discussing tech hub sustainability in Sub-Saharan Africa. The first will occur on Friday the 14th of November at 7AM EST or, 12PM GMT.

The first hangout will be hosted by Jon Gosier and Tayo Akinyemi and will focus on a discussion of sustainability and business models. You can join the hangout at the following link.

Several excellent introductory blogs have been posted:

Friday’s hangout will include participation by:

  • Mugethi Gitau, iHub, Kenya;
  • Yann Le Beux, CTIC, Senegal
  • Lukonga Lindunda, BongoHive, Zambia

Some of this discussion was a result of an observation based on our research teams visits to tech hubs and incubators across Africa. This should be a fascinating discussion. Please join us if you can!



Posted in Insights, News, Presentations, Research Projects | Tagged , , , , | Leave a comment

Networks and Military History


Toward the beginning of October I participated in the first Workshop on Digital Methods for Military History, held over two days at Northeastern University and funded by the National Endowment for the Humanities. Naturally, network science featured prominently in the schedule of events. The content of the workshop was divided into two major topics: network analysis and GIS/mapping. Each day proceeded in an orderly fashion: a round table of scholars discussing the topic of the day, presentations and explanation of projects, hands-on tutorials using a data analysis program, and finally a hack session to practice tutorial skills. The guest speakers included some of the most innovative people working in digital history today:

  • Jean Bauer, Associate Director of the Center for Digital Humanities at Princeton University. Focus: historical networks, early America, Project Quincy.
  • Scott Nesbit, assistant professor of digital humanities at the University of Georgia. Focus: digital maps, Visualizing Emancipation.
  • Alberto Giordano, Professor and Chair of the Department of Geography at Texas State University. Focus: mapping, geography of genocide, Holocaust Geographies Collaborative.
  • Micki Kaufman, fourth-year graduate student in US History at the Graduate Center of the City University of New York. Focus: network analysis of text, big data, Quantifying Kissinger.
  • David McClure, digital humanities software developer and interactive designer, late of the Scholars’ Lab at the University of Virginia. Focus: Neatline, digital mapmaking and curating.
  • Ed Triplett, advanced graduate student at the University of Virginia. Focus: ArcGIS, viewshed analysis, medieval Iberia.
  • Rob Warren, postdoctoral fellow at the Big Data Institute at Dalhousie University in Halifax. Focus: big data data-mining, semantic web, World War I.

As an historian with a long-standing interest in digital humanities, and with a lot of social network background outside Network Science, I was keen to be brought up to speed on the applicability and potential of Network Science (particularly) and ArcGIS to historical problems. And I was not disappointed. The interest level and energy were both high among forty-odd participants, and underscored just how new network analysis is for many historians (for current news and resources on the subject, I learned that is a great place to start). The most challenging aspect of Network Science, to me and it seemed to many others, was the methodological rigor involved in assembling raw data. “I had to learn how to make a proper database” was a refrain heard several times from the guest speakers, most of whom had not begun their careers in networks or mapping, but had to go through what might be called a methodological “OODA loop” to arrive at their current status. Not surprisingly, much of my evening time was spent rethinking, reworking, and (re)creating the databases for my own network research in medieval military and Civil War history.

The potential for Network Science to affect the study of military history really remains largely untapped, and the next edition of this NEH Workshop, should there be a sequel (as we all hoped there would be) will doubtless draw even greater interest and participation from military historians.


Posted in Conference/workshop | Tagged , , , , | Leave a comment

Newsblast Volume 4 Issue 11


In the current issue of the Network Science Center Newsblast Dr. John James discusses another aspect of the information sharing problem. He discusses Bitcoin’s solution to the ‘double spend’ or ‘Byzantine Generals’ problem and how this solution might be applied in other areas. To learn more read the article in this issue of the Network Science Center’s Newsblast, click here.

Posted in News, Research Projects | Tagged | Leave a comment