You gotta love this music video of Ed Sheeran leading the Sesame Street crew in a song about the differences between who children are at home versus school–and how to manage those unique spaces.

There’s a lesson in the song, too, for policy makers.

While school is a vital part of society, it alone cannot make up for the gaps in support children might experience at home or in their community. School has an essential role to play, but in coordination with other essential institutions.

We can use numbers to better understand literature. That’s right: there are fruitful ways for readers to deepen their analysis of books that are computational and quantitative. Doing so can be really intimidating though, evoking terms like text analysis, natural language processing, and data science.

There have to be better ways to create entry points for readers to explore literature quantitatively.

After exploring this idea for a few years now, I have developed what I think will be an accessible and enjoyable activity for readers and English teachers alike. Allow me to walk you through the process for using my customized Shakespearean spreadsheets–which I affectionately call “readsheets”–to create charts that quantitatively present how collections of words and themes develop throughout his plays.

OK. Let’s dive in.

Step 1: Count Every Word

We begin with this spreadsheet. It contains the frequency of every word used in Hamlet, excluding stop words (functional words like and, a, the, etc.)*. I like presenting the spreadsheet in long form, for reasons that will become clear in a moment. As a result, the column headers indicate the Act/Scene of the play.

Step 2: Identify Your Inquiry

In this example, I focus on using the spreadsheet during or after one has already read the play. It’s a way to analyze the text more deeply, to stoke text-based discussions and insights. (Though it is totally possible to use the spreadsheet as a pre-reading exercise as well.) In the case of Hamlet, it is not uncommon in high schools, for instance, to examine Shakespeare’s diction as it relates to death. For our purposes, let’s consider a version of such a prompt, but with a computational twist: Using both quantitative and qualitative data, tell me how Shakespeare portrays death in Hamlet.

Step 3: Create a Customized Table

Now that you know what you are interested in, it’s time to identify all of the Bard’s deathly diction AND to tally how many times each word is used in each scene. It sounds overwhelming, but really is quite simple. Go over to the first column and click on the filter drop-down. Above the list of words, you will see a search box. In it, type “death.”

In a split second, you will see all the words that contain “death” on the spreadsheet. This is worth playing with, because Google Sheets lets you select a few words and then search for others you can select as well. Copy that filtered list.

Next, paste the list of filtered words on the Custom tab of the spreadsheet. That’s where you are building your own customized table to visualize later. Finally, repeat the process for other words that you or students believe refer to death–directly or indirectly. In my sample, I include variations of the following words: death, sleep, and dream. It’s all about using the first tab to filter, and then copying interesting words onto the second tab. That’s the gist.

Once you have the tallies for individual words set, use the SUM function to create a tally of your collection of words. Essentially, you manually create a new data point that captures words-related-to-death-in-Hamlet.

Finally, now that you have your new tallies, you need to create a new table from which you will generate your graph. First, copy your SUM row and Paste Special below it, pasting Values Only to the new row. (This strips the formula, which would likely result in an error later.) Second, copy the top row of the table, which includes the Acts/Scenes, and Paste Special it below your table using the Transpose option. Do the same for the bottom Sum row. Rename the headers Scenes and Frequency (or whatever you like) and you should be set to chart! (This could also be done on a separate table, if that is preferable.)

Step 4: Make a Graph of Your Keywords

So, you have a customized data table with the sum of your keyword tallies. Now you can create a chart. It’s easy enough to do so right there in Google Sheets. Just highlight the new data set you created. Click on Insert and select Chart. For these data, an area chart tends to work really well. Once the basics look right, you can also play with the Chart Editor options, which is how I got different labels and colors. (Note: You can also copy your data set and paste it into other visualization tools online like this one.)

Step 5: Identify Insights and Inquiries from the Chart, and Re-read

Now that you can see the data visualized, what insights and inquiries surface for you? When I study the chart above, I wonder what is happening in Act 2, Scene 1. It’s one of only a handful of moments where the Bard’s deathly diction disappears. The scene is where Ophelia tells Polonius that she thinks something is wrong with Hamlet, highlighting the idea that Hamlet might not just be glum or eccentric, but mad. When I re-read the brief scene, I am drawn to this passage in which Ophelia says:

My lord, as I was sewing in my chamber,
Lord Hamlet, with his doublet all unbrac’d,
No hat upon his head, his stockings foul’d,
Ungart’red, and down-gyved to his ankle,
Pale as his shirt, his knees knocking each other,
And with a look so piteous in purport
As if he had been loosed out of hell
To speak of horrors, he comes before me.

Though she doesn’t speak of death here per se, her line “As if he had been loosed out of hell // To speak of horrors” resonates with me. It suggests that Hamlet re-appears as a phantom en route from hell. That is, his earthly behavior and mental state is one that transforms him into a tortured spirit. Though explicitly Shakespeare’s deathly diction is absent, the passage introduces a powerful concept: that the mind itself is the bridge between this world and the next. It’s like that line in Paradise Lost where Milton’s Satan declares “The mind can make a heaven of hell, a hell of heaven.” Hamlet’s mental state is the epicenter of the play, particularly as it relates to death. As I processed what the qualitative data (Ophelia’s excerpt) was suggesting, a new inquiry emerged: What is the correlation between Shakespeare’s language of death and his language of the mind?

I decided to go back into the data. This time, I created a new dataset for words associated with the mind, including: think, mind, brain, and mad. It looked like this:

Following the same steps as before, I created a new column for mind-related words and added it to my previous table. The result is a compelling visualization that offers some quantitative perspective on how Shakespeare portrays the relationship between death and the mind.

Were I teaching this text now or discussing it with my book club, I would love to explore the relationship between death and the mind further, especially in those few moments in the play where the language of the mind outshines the language of death–in 2.2 and 3.4.

Step 6: Try and Share with the World

Now that you have seen how I created a data visualization with Hamlet, it’s time to try one for yourself. I’ve made data sets available for some of my favorite Shakespearean plays! Here they are:

• Hamlet
• Romeo and Juliet
• Twelfth Night
• Midsummer Night’s Dream

And I’d love to see what you create. Please follow and tag me on social media (@tomliamlynch) and I’ll check out your creations. Now off thou go!

———–

* I create these “readsheets” for individual works of literature, ensuring that they are as accurate as possible. It requires previewing e-texts for boilerplate disclaimer language and transcription oddities, manually structuring the texts for analysis, verifying that the text mining program is reading the work correctly, and ensuring nothing gets lost in translation when saved as a spreadsheet. I’m exploring ways to create these for more works. If you have one you’d like to see, tell me on Twitter.

This quote comes from a thoughtful conversation I had with Chalkbeat NY about the city’s rollout of online learning in the wake of Covid-19. Specifically, it was in response to the city’s prohibition of schools using video conference provider Zoom. Here’s the clip:

Tom Liam Lynch, who runs the website InsideSchools and previously consulted with the city on education technology issues under the Bloomberg administration, applauded the move away from Zoom, given the security concerns.

Still, the disruption to schools demonstrates that the education department has lacked a coherent online education strategy for years, he said, and argues the department should appoint a deputy chancellor of digital learning.

“When you’re the largest school district in the country, all of this needs to be thought through before uttering a word about anything,” Lynch said. “If the city had a clearer digital learning plan, a lot of this heartache would have been avoided.”

Read the entire article by Alex Zimmerman over on Chalkbeat NY.

On the heels of 1.1 million students and 80,000 teachers being transitioned to remote learning, I penned this op-ed piece for The New York Daily News. Here’s how it starts:

I’m a firm believer that readers can use computational methods to uncover new insights into literature. As I’ve written about elsewhere, the quantitative can inform the qualitative, and vice versa. This goes for enjoying literature in whatever way you choose, whether you are a library card-carrying bibliophile, a beach reader, or an English teacher. I love exploring ways to computationally read texts. Not because numbers and graphs are the end in and of themselves. Hardly. Reading is about human beings better understanding themselves and the world around them. I just happen to believe that computers can also help us do that. In a recent experiment, turned to one of the most taught novels in high schools across America: The Great Gatsby.

When I think about F. Scott Fitzgerald’s most successful novel, I immediately fixate on the interplay between the story’s characters, as well as the significance of the car as amoral independence made manifest. What happens when we invite software to read The Great Gatsby with us? What new questions do we ask? What lines do we find ourselves rereading?

To explore this further, we start with data. I wrote a program that took every word in the novel, identified every chapter in which it appears, and then tallied those appearances. Since I’m interested in characters and the car, I focused on those. Well, it looks like this:

Numbers alone aren’t very helpful. But with some simple data visualizations, we can start to analyze the data and surface new questions. Let’s start by looking at the narrator Nick Caraway and the mysterious Jay Gatsby.

Since the novel is in the first person, we don’t get a lot of Nick referring to himself. But that just makes the use of his name in the book all the more interesting. In the first chapter both names are used comparably, but then Gatsby’s name seems to fall off between the second and third chapters while Nick is still setting up the plot. Why does the main character’s name disappear like that? An intriguing question, but one I deferred.

Next, let’s look at Daisy and Tom, a relationship that always leaves me feeling disturbed when I read the book.

Tom’s domineering presence is quantitatively on display. Look at how the frequency of his name begins fairly equally with Daisy’s, which is a phenomenon we saw with Nick and Gatsby. (There’s mounting evidence that Fitzgerald introduces characters in this way. A rereading of Chapter 1 might prove it further.) But then, watch Tom overshadow his wife in Chapter 2. And yet, by Chapter 4, we start to see Daisy’s name usage in the novel increasing as she is finding her path out of the abusive relationship into the arms of Gatsby. But this is something Tom won’t allow.  As we look at her name usage in the last chapter, it appears that she and Tom are somehow in sad balance again.

Now for the story’s narrative epicenter: Gatsby and Daisy.

In all the times I read and taught this novel, I don’t think I ever noticed just how Fitzgerald introduces Daisy to readers, far outshining Gatsby’s own name in the first and second chapters. Looking at the data, I’m also struck by how much more frequently Daisy’s name is used in Chapter 7. Gatsby’s name surges as well, of course, as the novel reaches its climax. But Daisy’s name more so. Perhaps the reason for her surge will come into focus if we add Tom to this area chart.

Whoa. Look at the sheer quantitative dominance of Tom Buchanan in Chapter 7. What’s more, observe the way Fitzgerald introduces Tom in Chapters 1 and 2, his name nearly always uttered somewhere. But much like his behavior in the book, the frequency of his name usage builds and brews and then blows up. Relatedly, the thinness of his name’s use in Chapters 8 and 9 might suggests that he, like a ravenous beast that gorges on his kill, rests when he is done. And like a beast on the hunt, Tom uses whatever advantage he can to succeed. In this case, a car. Let’s look now how the word ‘car’ factors into the story.

As you can see, the word appears consistently throughout the novel, with its usage never being great but always uttered somewhere. Then, we observe just how correlated the car is to the love triangle between Gatsby, Daisy, and Tom. A line graph shows the relationship even more cleanly.

My computational reading of the novel draws me to this passage in Chapter 7 I hadn’t noticed originally. Fitzgerald writes:

There is no confusion like the confusion of a simple mind, and as we drove away Tom was feeling the hot whips of panic. His wife and his mistress, until an hour ago secure and inviolate, were slipping precipitately from his control. Instinct made him step on the accelerator with the double purpose of overtaking Daisy and leaving Wilson behind, and we sped along toward Astoria at fifty miles an hour, until, among the spidery girders of the elevated, we came in sight of the easy-going blue coupé.

Fitzgerald might say that it is Tom’s wife and mistress “slipping precipitately” from his control, but Tom too is slipping from his own control. He experiences “hot whips of panic.” In life, some gain control by slowing down, breathing deeply. Others speed things up, like a child trying to find balance on a two-wheeled bicycle. Tom does the latter. He thinks speed will restore his control over life. Tom might be “of a simple mind,” but don’t mistake that for someone who doesn’t want to know precisely that everything–and everyone–is in its proper place. And he will go to vulgar depths to ensure that order is restored.

I’d love to hear what you see in the data. What questions emerge for you? What parts of the book do the data lead you to? And what new insights emerge as you reread the text with the numbers in mind?