Teaching the Digital ‘Abbāsids: An Introduction to Social Network Analysis

Sara Ann Knutson, University of British Columbia

An entry-level classroom activity for teaching and learning humanities and social science applications of introductory social network analysis. This activity can be used as a lesson plan or a tutorial handout for the classroom as well as an independent, self-guided learning resource.

Introduction

This article provides instructions for an interactive classroom activity that introduces learners to the data that Islamic (‘Abbāsid-era) coins can provide scholars in the humanities and social sciences and to the possibilities and limitations of social network analysis. The activity is designed as an introduction to learners who may be unfamiliar with network analysis tools and methods, databases, and analyzing data patterns using digital, open access software. The activity can also be applied to more specialized classrooms in which learners can interrogate an example of a digital tool that can help scholars examine data from material (archaeological) evidence. The article provides post-activity discussion and assignment questions that encourage critical evaluation of the possibilities and limitations of social network analysis in pedagogical settings. This work contributes to wider pedagogical and digital humanities scholarship by reflecting on what introductory data science and digital tools can offer for ethical pedagogies in humanities and social science classrooms, in which these methods are not often taught, and how teachers and learners can critically analyze data as non-neutral engagements of the past.

Project Background

This activity emerged from both my teaching of an Islamic history undergraduate course1 at the University of British Columbia (UBC) and my research on Islamic coins that began during my graduate studies at the University of California, Berkeley. During my first year as an Assistant Professor of Teaching at UBC, I designed an upper-level course, “The ‘Abbāsids: History, Heritage, and Memory” (now called “The Islamic Golden Age”), that addresses the past’s palpable legacy in the present. The course decenters Europe in engaging the premodern Islamic past—a burden that otherwise often falls unevenly on history teachers and scholars of the contemporary Middle East.2 The course continues to appeal to students from diverse positionalities, learning styles, and disciplinary backgrounds. Former students have communicated that the course provides an all too rare space for those who culturally identify with the Middle East to feel seen on campus and, for students who do not culturally identify with the region, to become more curious and invested in anti-colonial efforts and problematizing harmful biases.

This digital activity appears in the ‘Abbāsids course during a thematic week titled “The ‘Digital’ Islamic World.” During this week, I guide students through this activity, which uses data from a sample of approximately 1,500 archaeologically-uncovered Islamic coins. Based on their learning of ‘Abbāsid trade and global connections in the previous week, students learn that these identified coins hardly scratch the surface of the number of Islamic coins deposits uncovered across Afro-Eurasia (Kovalev and Kaelin 2007). I provide this initial historical context before students engage with data in order to first communicate that Islamic coins are material traces of the Islamic World, from which they originate. It therefore does not naturally follow that these coins must necessarily be interpreted as “exotic” silver that aggrandizes medieval Europe’s trade connections. Recognizing that the dataset is not a representative sample of the historical phenomenon of global Islamic coin circulation during the ‘Abbāsid period, the goal of the digital activity is rather to introduce students to the possibilities and limitations of network data to engage the past.

Prior to beginning the digital activity, students review examples of digital humanities projects (Wrisley 2015, 2016; Muhanna 2016; Romanov 2018) and evaluate how scholars use digital methods to study historical evidence, the challenges of and the issues at stake in using various digital methods, and the claims and assumptions that researchers make in their scholarship. Students learn that digital tools and data are not neutral—like texts, materials, images, and other evidence, they require interpretation. And as Hussein Keshani (2012) similarly explains, digital technologies do not simply provide tools for archiving or exchanging information; these tools are themselves also analytical instruments and communication platforms that are embedded with under-scrutinized theoretical assumptions. During the activity, students navigate the software Gephi on their personal laptop computers, which they bring to class. As open-access software, Gephi is free to all users and operates on the three most common computer operating systems, Windows, macOS, and Linux, which means that most students with a computer will have a compatible device for the activity. Students can alternatively partner with a classmate with a compatible computer for this activity in the event that not all students have access to a computer in the classroom. Each learner in this activity produces two network graphs as I guide them through the sequential steps, pausing at various times for check-in and discussion. Once the class has successfully created the network visualizations, we conclude the classroom activity with discussion and feedback.

The Social Network Analysis Activity

Introduction

Social network analysis (SNA) is a methodological approach that assembles and organizes relational data, represents the data as a network, and analyzes the resulting network (also called a graph) for potential patterns of connection.

A network consists of two basic elements: points and lines. Depending on your discipline, points may be alternatively called vertices, nodes, sites, or actors, and you may see lines alternatively called edges, arcs, links, bonds, ties, or relations. For the purposes of this activity, we will call them nodes and edges.

There are two basic types of networks: directed and undirected. In directed networks (sometimes called communication networks), the connection between nodes has a certain direction. For example, when you send a letter to a friend, that letter travels in a certain direction, from you to your friend. Or when you follow a celebrity on social media, that does not necessarily mean that the celebrity will follow you back—in such a case, your “follow” goes in one direction. In undirected networks (sometimes called friendship networks), the connection between nodes goes in both directions. As the alternative name implies, an example of an undirected network would be a friendship—a connection between two people that goes in both directions (at least ideally!).

In this activity, we will use a free, open-source network analysis and visualization software called Gephi. For our dataset, we will examine data on archaeologically-uncovered Islamic (‘Abbāsid) silver coins, called dirhams. This data derives from first-hand data collection in museums, archives, and online databases (including the Hungarian National Museum, Budapest; the Money Museum, Zurich; National Museum Warsaw; the Ashmolean Museum, Oxford; National Numismatic Collection, Amsterdam; the Portable Antiquities Scheme, UK; Statens Historiska Museum, Stockholm) and from published archaeological reports (CNS- Corpus Nummorum Saeculorum: Catalogue of Coins from the Viking Age found in Sweden, developed by the Stockholm Numismatic Institute, volumes 1.1 Gotland, Akeback Atlingbo; 1.2 Gotland, Bal Buttle). For an introductory article to the wider historical phenomenon of circulating Islamic silver coinage (including during the ‘Abbāsid period), see Kovalev and Kaelin (2007).

An important note on the dataset for this activity

This dataset is intended for educational purposes only; the dataset has been modified for the purposes of this SNA activity and does not necessarily contain all information available on the coins that are housed in these archival collections nor on the coins indicated in the published catalogues. The archival and published sources used here are also by no means the only repositories for data on surviving ‘Abbāsid coins. This dataset therefore does not provide a representative, generalizable sample for research on the wider historical phenomenon of circulating Islamic coinage in Afro-Eurasia.

Setting up the software

A) Download Gephi

Begin by downloading Gephi and opening the software. At the time of writing, you can access the software at the following link: https://gephi.org/users/download or otherwise by accessing the Gephi website. Note that Gephi is available for macOS, Windows, and Linux systems. (For Mac users, if Gephi does not open after downloading the software, check your security settings [Apple menu > System Settings > Privacy & Security. Under Security, select “open anyway” under security box indicating that Gephi was blocked from use].)

B) Download dataset

Navigate to https://github.com/SAKnutson/SNADemo_Islamic_Coins.git and download the ZIP folder labeled SNA Demo of Islamic Coins (click on this ZIP folder name, navigate to the button labeled Download raw files). The ZIP file contains two folders (labeled Part 1, without edges and Part 2, with edges) with three total Excel files. Unzip the folder. It is recommended that you move these files to your desktop or to an alternative location on your computer that is easily accessible.

Reviewing the database

Before working with any digital software, it is good practice to begin by reviewing the dataset itself. Navigate to the “Part 1” folder and open the enclosed Excel file. The original Excel file was created in .xlsx format. However, in order to prepare for SNA work in Gephi, this .xlsx file was converted to .csv (comma separated values) format.

Discussion suggestions

As a group or individually, start by considering:

• What kinds of information are available in the database?
• What do each of the data categories (i.e., each Excel column) mean?
• Why might it be important that every entry (representing one coin) have a unique ID number?
• Hypothesize: why do you think that there are non-existing data (i.e., blank entries in the spreadsheet) in the case of some coins?

Pro tip

When you work with any digital humanities project, it is often good practice to keep an original file of your database/dataset with a timestamp in the file name and to make subsequent versions with their own timestamp as well. (For example: DATABASE_15_July_2023.) Then, any time you need to make modifications to the original database (such as preparing the database specifically for SNA software), first duplicate the database into a separate file and make any modifications to the second version of the database. Save this file with a new name. (For example: DATABASE_SNA_20_AUG_2023.)

Before you finish reviewing the database and making some mental notes of what you notice about the information available (and which information is not available in the database), make sure that the lat/long (latitude/longitude) coordinate data is in decimal format, NOT degrees and decimal minutes (the decimal format is important for working with Gephi).

Part 1: Creating the first network

To create the first network, begin by closing the Excel database files if you wish and opening Gephi. If the program prompts you to open a file/previous project, you may simply close that prompted window.

A) Plugin setup

For this activity, we will use some specific Gephi plugins that are not automatically installed by downloading Gephi. Navigate to the top menu in Gephi and select: Tools > Plugins

Click on the Available Plugins tab. In the search bar on the right side of the window, type geolayout (make sure this is one word).

Check the left-side box under the Install column for geolayout and click the Install button at the bottom. (Follow the prompts for the Plugin installer and select Restart Later and Finish.)

Complete the installation process for one more plug-in. Navigate to the search bar and type map of countries (these are three separate words).

Check the left-side box under the Install column for map of countries and click the Install button at the bottom. (Follow the prompts for the Plugin installer and select Restart Now and Finish.)

B) Importing the spreadsheet

Open Gephi and from the top menu, select File > Import Spreadsheet… (Navigate to your computer desktop or the location where you saved the database files and select the first Excel file, titled Part 1_Knutson SNA demo.)

A new window will appear. In this import window, you must ensure the following:

• The Separator drop-down option should be set to Comma
• The Import as drop-down option should be set to Nodes table
• The Charset drop-down option can be left at UTF-8

Select Next > . This will take you to the second and final Import settings page.

There are different ways of categorizing data in Gephi. String refers to any non-numerical value (for instance, the name of an ‘Abbāsid caliph is not a number; therefore it is categorized as a string). A double is any numeric variables consisting of numbers with decimal points (for instance, the decimal format of latitude and longitude coordinates). An integer applies here to dates.

Note that Gephi cannot handle any non-standard formats for dates, such as years. This is often a problem for anyone working with the conversion between Islamic and Gregorian years. For example, an Islamic coin minted in the Islamic (Hijri) year AH 350 converts to 961/962 CE in the Gregorian calendar. Gephi, however, would not be able to handle a data entry of “961–962” as an integer; the software can only manage a single year (such as “350”, when referring to the year or “961,” when referring to the Gregorian year).

If students in the class are unfamiliar with the conversion between Islamic and Gregorian calendars, this is a good moment to pause and briefly discuss how a single year in the Islamic calendar (without further information about the specific month or date, which Islamic coins do not provide) might translate to an ambiguous year range.

This is also a good opportunity in the classroom to discuss a common best practice in digital and data work: in cases where software capabilities (and limitations) require you to make decisions about how to represent data in your database, it is good practice to keep a written record of the decisions you made when cleaning up data in secondary versions of your database.

In this second Import settings window, ensure the following:

• The Time Representation drop-down option should be set to Intervals
• The Museum Collection drop-down option should be set to String
• The Source drop-down option should be set to String
• The Target drop-down option should be set to String
• The Mint Authority drop-down option should be set to String
• The Location of Mint drop-down option should be set to String
• The lat drop-down option should be set to Double
• The long drop-down option should be set to Double
• The Findspot (City) drop-down option should be set to String
• The Findspot (Country) drop-down option should be set to String
• The Find_lat drop-down option should be set to Double
• The Find_long drop-down option should be set to Double

Once you have selected these options, click Finish.

If you see no issues reported, congratulations! You are ready to work with the network! If you see a list of errors at this stage, it may be because you forgot to change one or more of the drop-down options to the correct category. Remember to import this database file as Nodes table, NOT “Edges table”! To try again, select Cancel and repeat the steps listed in part B of this section.

C) Organizing and examining the network

At this point, you will see an unorganized network (graph). Note that each node here represents an individual Islamic coin that was uncovered archaeologically (most often within the past two centuries). The nodes here are not yet represented in any geographic space. We will get to that soon.

Before we organize and analyze the network, let’s briefly explore some basic functions of Gephi…

Basic functions of Gephi

When guiding students who have never worked with Gephi before, it is important to allocate time in the classroom to ensuring that students gain a sense of the basic windows and tools in the software before continuing on with the activity. Students often have questions about the difference between the Overview and the Data Laboratory windows and how practitioners might work between these two tabs. The next few steps in this activity are therefore included to scaffold some basic, foundational knowledge of how to navigate the basic tools of Gephi before students build on this knowledge and continue working with the data in this activity.

(For instructors or students who have prior experience with Gephi, you are also welcome to skip ahead to “Now let’s change the appearance of the network.”)

Click on (a) the Data Laboratory tab to view the database within Gephi.

Scroll through the data laboratory and click on any node of your choice so that the entire database line is highlighted. (Remember, each node represents a unique Islamic coin.)

Now, right-click on your selected node. Click Select on overview.

Then, navigate to the top tabs and select (b) Overview to switch to the main screen with the network. You should now be able to see the node you selected in the data laboratory in its current position in the network.

Select the (c) direct selection tool from the left graph sidebar.

Now, let’s try that process again in the opposite direction. Start by getting a sense for how to zoom in and out of the network in the Overview tab. You can zoom in and out by scrolling with your mouse or trackpad, or select (d) center on graph from the left-side graph toolbar.

Find a node from the network that you would like to select, hover over the node and right-click. Choose Select in data laboratory. Now navigate to the Data Laboratory tab to see your selected node in the database (it will be highlighted). This function is useful for identifying a particular node in the network graph.

Network appearance

Now let’s change the appearance of the network. In the Overview tab, navigate to the (e) Appearance section (usually in the left-hand side). Under the Appearance section, select the Nodes tab, select Partition, and from the drop-down options for choose an attribute, select mint authority. The mint authority is the ‘Abbāsid caliph under whose reign a given coin was created. Select Apply at the bottom of this section to apply this appearance category to the network.

This function already shows you some useful information about the coins available in this database, namely which caliphs are represented the most in this particular dataset. (Please note that this dataset is NOT representative of the vast wider phenomenon of Islamic coins that were minted during the ‘Abbāsid era, nor the vast number of coins that survive archaeologically!)

Now we are ready to see the network represented in geographic space. To do this, move to the section below Appearance, labeled Layout. (If the Layout tab is not already visible or open for the user, you can access this tab by navigating to Window > Layout.)

From the drop-down options, select Geo Layout.

In the settings under Geo Layout, please ensure the following:

• Scale: 1000.0
• Latitude: lat
• Longitude: long
• Projection: Transverse Mercator
• Center: [uncheck]

Once you have selected those settings, click Run.

A few issues can emerge. Sometimes Gephi appears to run indefinitely after the user initiates the “run” process. If the user clicks “run” and they can see the subsequent reorganization of nodes, they have successfully completed this process. In cases where Gephi continues to run after this point (note that this is common!), simply select “stop” and move on to the next step.

In worst case scenarios, such as if Gephi seems to have become frozen, you may need to close Gephi, restart the software, and try again. If this happens, do not despair—Gephi can be a little unwieldy sometimes!

If we were preparing this network visualization for published research, we might consider making a copied version of the database and, in the modified version, removing the nodes that do not contain spatial information—assuming that kind of information was important for our analysis and/or visualization. Note that if we do that, it is very important to indicate to the reader that the resulting network visualization does not include coins for which no available spatial information exists. It is important to not deliberately mislead potential audiences who might view our network visualization!

At this point, the network still does not look like much, but you will hopefully notice some changes.

Next, return to the Layout section and from the drop-down options, select Map of Countries. (Please note that in order to see the distribution of Islamic coins on a map, this process cannot be done in reverse order—we must run the Geo Layout plugin before Map of Countries. To redo this process, you must exit Gephi and reopen the software. Unfortunately, there is no “undo” option in Gephi.)

In the settings under Map of Countries, please ensure the following:

• Country: world
• Subregion: no subregion
• Region: no region
• District: no district
• Scale: 1000.0
• Weight: 1.0
• Projection: Transverse Mercator (the projection should match the one selected for the Geo Layout plugin)
• Center: [uncheck]

Once you have selected these settings, click Run.

Take some time to examine and move around the map on your own. The network visualization you have successfully created (see Figure 9) shows the locations in the Islamic World where the specific coins in the dataset were minted. Coins that do not have this information are, again, represented as nodes along the black line at the bottom of the network graph.

This is a useful time to mention to students that while it is fun to see the result of the network visualization in Gephi, the visualization is not always the ultimate goal of researchers. Network visualizations like this one can also be useful research tools in order to double-check the data that is available in the database and ensure that there are no obvious errors or mistakes.

This would also be a good time to double-check the data in our database…

Using what you have learned so far in this activity, click on any node of your choice in the Overview tab and examine the same node in the Data Laboratory. Compare the information in the data laboratory to what you see in the network graph. For example: does the city information listed in the Data Laboratory tab seem to reasonably fit where it is located on the map in the Overview? (This is worth checking, because it is quite easy to accidentally assign the wrong geographic coordinate information.)

This is a great moment to ask students: “As you select a node from the network in the overview tab and then navigate to the data laboratory tab, you may notice that multiple nodes are highlighted in the database view. What does this mean?” Hopefully, you’ll get an answer noting that multiple nodes being highlighted here means that multiple coins in the network visualization are represented by a single node! This is a very important point to keep in mind when interpreting network visualizations produced in Gephi—any node in this network visualization might represent a single coin or many coins. This is one limitation of this particular network visualization.)

Congratulations! If you have made it this far in the activity, you have successfully made a network graph in Gephi. As you may have noticed, this is a nodes-only graph (there are no edges—yet!).

When you are finished examining the network, take a screenshot of your graph and label the file: Nodes Graph_Gephi.

You may save your project in Gephi if you wish. When you are ready, exit the software.

Part 2: Creating the second network

To create the second network, begin by (re)opening Gephi. If the program prompts you to open a file/previous project, you may simply close that prompted window.

A) Importing the spreadsheet

From the top menu of Gephi, select File > Import Spreadsheet… (Navigate to your computer desktop or the location where you saved the database files, open the Part 2, with edges folder and select the excel file titled Part 2 NODES_Knutson SNA demo). This is the database for the network nodes.

A new window will appear. In this import window, you must ensure the following:

• The Separator drop-down option should be set to Comma
• The Import as drop-down option should be set to Nodes table
• The Charset drop-down option can be left at UTF-8.

Select Next > . This will take you to the second and final Import settings page.

In this second Import settings window, ensure the following:

• The Time Representation drop-down option should be set to Intervals
• The Museum Collection drop-down option should be set to String
• The Source drop-down option should be set to String
• The Target drop-down option should be set to String
• The Mint Authority drop-down option should be set to String
• The Location of Mint drop-down option should be set to String
• The lat drop-down option should be set to Double
• The long drop-down option should be set to Double
• The Findspot (City) drop-down option should be set to String
• The Findspot (Country) drop-down option should be set to String
• The Find_lat drop-down option should be set to Double
• The Find_long drop-down option should be set to Double

Once you have selected these options, click Finish.

If you see no issues reported, congratulations! You are ready to work with the network! Click OK.

B) Organizing and examining the network (nodes)

Same as before, we should check the data laboratory and briefly review the dataset. Note that coins that contain available information about both their location of mint AND their archaeological findspot will be represented twice in this database file. (If at any point the data laboratory does not show up for some reason, you can easily find it again by navigating to Window > Data Table.)

Back in Overview, you can change the appearance according to whichever category you prefer. (Select, for instance: Nodes tab > Partition tab > mint authority from drop-down options)

Now again, we are ready to change the network layout. Navigate to the Layout section in Overview.

First select Geo Layout from the drop-down options.

In the settings under Geo Layout, please ensure the following:

• Scale: 1000.0
• Latitude: lat
• Longitude: long
• Projection: Transverse Mercator
• Center: [uncheck]
• Looping: [check]

Once you have selected those settings, click Run.

Note that this process may take some time. You may notice a blue bar at the bottom of the window while this process is operating. If this takes several minutes, try cancelling the process by clicking on the [X] box next to the moving blue bar (or click the Stop button where the Run button is normally located).

As soon as you see the reorganization of nodes similar to the image in Figure 10, you have successfully completed the process and you can click on the Stop button if this process does not stop automatically.

Next, return to the Layout section and from the drop-down options, select Map of Countries.

In the settings under Map of Countries, please ensure the following:

• Country: world
• Subregion: no subregion
• Region: no region
• District: no district
• Scale: 1000.0
• Weight: 1.0
• Projection: Transverse Mercator (the projection should match the one selected for the Geo Layout plugin)
• Center: [uncheck]

Once you have selected these settings, click Run.

The resulting network graph you have successfully created represents the coins’ city of origin (the mint center) AND the location where they were uncovered archaeologically, many centuries later. This means that each coin is potentially represented with two nodes in the network graph (we say potentially, because information on the mint center and archaeological findspot are not always available for every coin). Coins that have neither of these two pieces of information are again represented by nodes found along the bottom of the network graph.

What we now need are edges that connect these nodes and help represent where a given Islamic coin originated and where it was uncovered archaeologically…

C) Importing the edges and organizing the network graph

To import the edges into this network graph, select File > Import Spreadsheet…

Navigate to your computer desktop or the location where you saved the database files and select the third Excel file, titled Part 2 EDGES_Knutson SNA demo. This is the database for the network edges.

A new window will appear. In this import window, you must ensure the following:

• The Separator drop-down option should be set to Comma
• The Import as drop-down option should be set to Edges table
• The Charset drop-down option can be left at UTF-8.

Select Next > . This will take you to the second and final Import settings page.

In this second Import settings window, ensure the following:

• The Time Representation drop-down option should be set to Intervals
• The imported columns likely cannot be changed, but Source and Target are most likely already checked

Once you have selected these options, click Finish.

Hopefully, you will see no issues reported. There are a few more items to check on this page, however. First, make sure that directed is selected from the drop-down option under Graph Type. This can be a great time to ask students why, based on the information about this network graph provided above, it is a “directed” network (and not an “undirected” network).

Second, navigate to Edges merge strategy (you may need to click more options if you do not see this option). Make sure that don’t merge is selected.

Finally (and importantly!): select append to existing workspace. (We are working on the same network and do not want to create a separate workspace.) Click OK.

At this point, your network graph should look something like a crazy rainbow! That particular feature of the visualization is caused primarily by the fact that there are coins in the database for which the mint location is unknown, but the archaeological findspot is known.

.

Take a moment to examine and move around this map on your own and try to take note of any particular patterns that you see based on this visualization.

I use this opportunity to allow students to navigate around the network graph until they encounter something unusual—an Islamic coin represented in North America (see Figure 13)—meaning either means that there is a mistake in the database or that one Islamic coin travelled across the Atlantic by itself! We usually discuss explanatory historical and data-based possibilities for this apparent anomaly. The correct answer here is that this observation indicates a mistake in the database—for this particular coin, the latitude and longitude information were erroneously switched. This is an example of how data visualizations are not always meant to be simply the “final product” of digital humanities work. In this case, the visualization helps us to quickly spot a mistake in the database!)

Whenever you are finished examining the network, take a screenshot of your graph and label the file Edges Graph_Gephi.

You may save your project in Gephi if you wish. When you are ready, exit the software.

Synthesizing Our Learning and Further Discussion

A) Post-activity discussion and/or assignment questions

• Based on the two graphs that you created in this SNA activity, brainstorm and discuss: what patterns did you notice based on completing this activity? How might we use these graphs (and the information they represent) in order to better understand the ‘Abbāsid past? What kinds of research questions might you pose that these current data and network visualizations (or future versions of these) might help us answer?
• What can we NOT observe from this kind of dataset and/or network visualization? What kinds of biases might be influencing the data and the network visualization?
• Based on your participation in this activity, what might be some best practices for managing databases that support social network analysis or for developing data that supports social network analysis?
• What can we learn from multiple lines of evidence from the ‘Abbāsid past, such as archaeologically uncovered coins? Textual sources? In what ways might we use these lines of evidence together?
• One of the important applications of this activity is the importance of placing data and digital humanities work in their proper historical and cultural context. Data and digital tools are not neutral—they require interpretation and decision-making processes. Given that point, how might we ethically use data like that explored in this activity to better support contemporary communities, including those who culturally identify with the Middle East and North Africa (MENA) region for whom Islamic coins and other materials are important to their cultural heritage?

B) List of terms introduced in the activity

Directed networks

Edges

Gephi

Graph

Integer

Lines

Network

Nodes

Points

Social network analysis (SNA)

String

Undirected networks

Reflections on the Activity

I designed this activity with some core pedagogical considerations and ethical values in mind. Firstly, my work is based on pedagogical scholarship and practices that reconceive teachers and students as active coproducers of knowledge,3 a pedagogical principle that aligns well with core values of openness and collaboration in the field of digital pedagogy (Davis et al. 2020). As a result, this activity is the product of layers of collaborative student input, engagement, and feedback. The database for this activity developed partly alongside two student researchers in respective university programs at the University of California, Berkeley and the University of British Columbia that offer students paid or for-course-credit experiential learning opportunities to work alongside faculty. Such programs most directly align with ethical pedagogies when faculty members refuse student labor that solely benefits themselves or institutions and instead appropriately position students as coproducers of knowledge with independent expertise and provide opportunities for students to participate and help shape the project according to their own specific goals, interests, and values.

Following our work on the database, this “Digital ‘Abbāsids” pedagogical activity first emerged as an informal guide that I designed to support my first student research assistant who expressed interest in acquiring introductory skills in social network analysis. After that mutual learning experience, I integrated this activity into my undergraduate history classroom as an opportunity to assess how digital tools like social network analysis might enhance opportunities for teachers and students to co-construct knowledge and resist expectations that learning happens only through passively observing or memorizing historical “facts.”

Based on my teaching experiences, I found that this activity encourages students to co-create knowledge by interacting with the data themselves, and I celebrate that they inevitably identify different patterns and ask different historical questions than I do, informed by their own positionalities, experiences, and interests. In anonymous feedback, students in the course indicated that they really enjoyed this activity as an interactive way of engaging the course content. One student wrote that “it was very beneficial to learn about the tools used in modern historical research,” while another indicated that the activity was “definitively informative,” and they suggested integrating more primary sources in that week in order to give additional historical context to some of the patterns that they saw in this activity. This insightful suggestion was then incorporated into subsequent iterations of the course design (see Knutson, Shaheen, and Izadi 2024).

As a second pedagogical consideration in designing this activity, I join digital pedagogy scholars who argue that digital tools are especially useful for helping students to understand their relationship to knowledge production (see, for instance, Risam 2018; Croxhall and Jakacki 2023). This learning about knowledge production works on at least two levels here. First, I designed this activity with the intention of highlighting the process of working with data and digital tools in order for students to each produce their own knowledge and engagements of the ‘Abbāsid past. Not least, this activity incorporates active discussion, reflection, and troubleshooting throughout the process of producing two network graphs. Following Battershill and Ross (2022), participants in this activity are encouraged to embrace the “unforeseen, the process-oriented, the accidental, and the contingent” (6) and to recognize that any “mistakes,” “uncertainties,” or otherwise “messiness” encountered in this activity are not weaknesses, but rather part of the active learning process of constructing knowledge (see also Collier and Ross 2017). On the second level, one learning objective of this activity is for students to better appreciate the analytical considerations involved in how knowledge producers assemble, use, and evaluate evidence, including data.

Based on my teaching of this activity to generate student learning of academic knowledge production, I have come to realize that digital tools and data science applications have much to offer humanities and social science classrooms, especially courses in which these methods are not frequently taught. I developed and integrated this “Digital ‘Abbāsids” activity into such a course where students are traditionally exposed to the primary source analysis of historical texts. In my course design, students indeed engage primary source analysis, but they also learn archaeological approaches to analyzing material evidence and, as readers know well by now, digital tools and data approaches. In other words, “The Islamic Golden Age” is not intensively nor exclusively a digital humanities course, but rather one that introduces students to digital tools and intersecting scholarly debates in the digital humanities and history, in addition to other methodologies. My point here is that introductory data science and digital tools do not simply belong in specialized digital humanities courses; these tools can have a profound, or perhaps even the greatest, impact in humanities and social science classrooms as innovative and integrative “small teaching” approaches.4

Finally, my activity design is informed by “ethical pedagogies”5—that is to say, practices of teaching and learning that deliberately center the processes of human inquiry and knowledge production, their nonneutral impacts on past, present, and future people, and the values of equity, empathy, and care. Ravynn K. Stringfield (2021) has expressed ethical concerns for digital humanities projects that exist simply because they can, without regard for their potentially harmful impacts. She observes that Black digital humanists’ projects, alternatively, “often center humanity and approach digital tools with an ethos of care” (477). Stringfield’s powerful point implicates (digital) pedagogy as well. Following my definition of ethical pedagogy above and the arguments of Stringfield, Lasmana (2021), and other contributors in Kim and Koh’s (2021) Alternative Historiographies of the Digital Humanities, ethical digital pedagogies can include practices (among others) which interrogate the limits of data and digital representation, encourage “polyvocality” (Lasmana 2021, 320), destabilize the presumed neutrality of data and digital tools, question the epistemic norms underpinning the construction of “academic” knowledge, and center values of social justice and equity.

My personal teaching of ‘Abbāsid History, including this activity, involve anti-Islamophobic, antiracist, and anticolonial values as well as decentering Europe and recognizing contemporary practices of engaging the Islamic past that academia often fails to value or consider legitimate. Therefore, in interactive engagements of digital tools and data such as this one, I worry that users may divorce the data from its proper historical and human context and center, rather than destabilize, Eurocentric, Islamophobic attitudes towards the Islamic past. In this context, I am inspired by Pheroze Unwalla’s (2021, 2022) practices of critical hope and emotion(ality) in and beyond the Middle East history classroom and Dorothy Kim’s (2021) and Loren Guay’s (2023) respective analyses of (dis)comfort in academic research and academic labor. Reflecting on these important pedagogical contributions, all of which encourage resistance to unethical practices and the centering, not avoidance, of discomfort, I choose to be transparent to students about my concerns, and these moments have resulted in generative classroom conversations that communicate to students the real stakes involved in studying the past.

Similarly, I have intended this activity as a pedagogical exercise that foregrounds the impermanent, emergent nature of the dataset rather than situating it in the supposed “objective” past for “objective” research aims. There is little “objective” about universalist claims to heritage, data, and history that mean something to students and their relationship to their pasts, presents, and futures.6 Therefore, one of the most important questions that we can ask as students and teachers in pedagogical settings that utilize interactive and digital technologies and data is “Who might this source or dataset serve?” and “When and how can we best and appropriately use these tools and data to engage the past in ethical ways?” No interactive activity is immune to these core questions.

Acknowledgments

Archival research for some data used in this pedagogical activity was supported with funding from the Swedish Women’s Educational Association (SWEA) San Francisco (2020), a UC Berkeley Center for Race and Gender student research grant (2021), and a Social Science Research Council (SSRC) Mellon International Dissertation Research Fellowship (IDRF) (2021). The database was partly developed with the research assistance of Yuxin Zhao at the University of California, Berkeley (funded by the Undergraduate Research Apprentice Program) and a student at the University of British Columbia (funded by a UBC History student research grant) who wishes their name to remain anonymous. I am very grateful to both of them for their respective contributions to the database. My reflections in this article emerged from my presentations at the 2022 Middle East Studies Association (MESA) Conference and the 2023 Digital Pedagogy Institute, and I appreciate the audience feedback at these venues, which further informed this project. Finally, I am grateful to my former UBC students in the course “The ‘Abbāsids: History, Heritage, and Memory” (Spring 2023) for their valuable feedback on an earlier version of this activity.

Notes

1. For the published syllabus for this course, please see Knutson, Shaheen, and Izadi (2024).↑

2. But see, for instance, Rachel Schine’s (2022) work on teaching race with premodern Arabic texts and her discussion of the complex dynamics of teaching translated premodern sources. ↑

3. This theoretical foundation is informed by Lave and Wenger’s (1991) concept of legitimate peripheral participation. For an earlier example of how I incorporated this into my teaching, see Knutson (2018). ↑

4. I use James M. Lang’s (2021) term here to gesture to the kinds of pedagogical approaches that “seek to spark positive change in higher education through small but powerful modifications to our course design and teaching practices” (4). This approach aligns well with Ryan Cordell’s (2016) recommendations for incorporating digital pedagogy into the undergraduate classroom by starting small, integrating digital tools into broader disciplinary debates and inquiry, and destabilizing notions of the Digital Humanities as something profoundly “new” to our undergraduate students. ↑

5. See also Jones (2022). ↑

6. For a wonderful demonstration of how research, not least in science, is fundamentally a human endeavor and therefore subject to human assumptions and power-dynamics, see Prescod-Weinstein (2021). ↑

Appendix: Further Reading and Resources

Professional networks

Islamicate Digital Humanities Network, https://idhn.org/

The ISHN website features an extensive “resources” page with online resources and tools for teachers, students, and researchers who wish to pursue more specialized work in Arabic and Persian natural language processing (NLP). The page also includes online databases and resources to digitized Islamic(ate) primary sources (including texts and audio sources).

Additional pedagogical resources

• GEPHI Demo for a European letters dataset. Grandjean, Martin. 2015. “GEPHI- Introduction to network analysis and visualization.” http://www.martingrandjean.ch/gephi-introduction/
• GEPHI Demo of characters in the novel Les Miserables. University of British Columbia Library. https://ubc-library-rc.github.io/gephi-palladio/content/05_Gephi_hands_on_activity.html
• Introduction to Social Network Analysis (free online course taught by Dr. Martin Hilbert, UC Davis, provided by Coursera): https://www.coursera.org/learn/social-network-analysis
• Overview of Social Network Analysis (developed mainly for Public Health audiences) with further readings, software, websites, and courses, Columbia University: https://www.publichealth.columbia.edu/research/population-health-methods/social-network-analysis#:~:text=Social%20Network%20analysis%20is%20the,%2C%20groups%20and%2For%20organizations

This entry is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license.