Designing a project for my digital humanities class (draft 0)

Topics/tags: CSC 151, Racket, teaching, digital humanities, long, rambly

For about as long as we’ve been teaching a themed version of CSC 151, we’ve included an end-of-semester project. The project serves a variety of purposes. It gives students a sense of ownership of their work; people feel differently about projects they design than projects we design. It expands students’ understanding of possibilities; seeing what other students undertake reveals very different approaches to computing. It reminds them that they need more than technical skills to succeed; because the project requires a written report, a short presentation, and an even shorter question and answer session, students must draw upon and build on those skills.

Looking back on the old MediaScheme version of CSC 151, it’s clear that the project evolved over the multiple semesters we offered it. Our instructions got more specific. We added examples [1]. We wrote a rubric [2]. We revised the rubric. We tried different timings and approaches to classroom time. We ended up with a pretty good structure, if I do say so myself.

Two things about the MediaScheme project will be hard to translate to other projects. First, we could choose a very consistent structure: Students had to write a procedure, series, that took three arguments: a width, a height, and a number between 0 and 999. Second, we told them that a member of the studio art faculty would assess their work. The structure meant that we knew students would be challenged to make write algorithms that stretched and scaled the image appropriately, thereby exercising their algorithmic muscles, as it were. The external audience meant that students needed to concern themselves with more than algorithms; they had to think about what would make an image (or series of images) compelling. Translating those characteristics to a new project will prove challenging.

There are ways that I could restrict the project. I could limit the data set that students are using; for example, I might require that they use our digitized version of the Scarlet and Black or to the Statistical Accounts of Scotland [3]. I could limit the kind of project they do, such as requiring everyone to do a language generation project or a mapping project.

On the other hand, there are some exciting things about moving to a more open project model. It certainly gives students more opportunities to explore their passions. It provides a broader range of possibilities [4]. And I’ll probably have to worry a bit less about repetition from year to year [5].

One of the downsides of having a more open-ended project is that it’s much harder to predict workload. Students may also spend more time on background work, such as identifying an appropriate data set. There’s also a challenge of being able to grade the projects uniformly. However, a good rubric can help with that.

The other strength of the MediaScheme project was that they knew they would have an artist look at their work. I could give them an audience, of sorts, telling them that we will invite one of the digital humanists on campus to critique their work. That’s an issue to consider for the future.

For this draft, I’m going to use the open-ended project model. I may add a short mention of the audience. Let’s see what I can come up with. I have as resources the latest version of the MediaScheme project description and the latest version of the new, and not nearly as polished, Data Science project description.

Preliminaries

Summary: You will design and propose a novel project in the digital humanities; develop, implement, and apply appropriate algorithms; and present your findings in both written and oral form.

Purposes: The project gives you the opportunity to explore some aspect of the digital humanities in more depth and to do work that is novel and creative. We also intend that it will encourage you to reflect more purposefully on the design of algorithms.

Collaboration: I encourage you to work in groups of size three. (In class, we will discuss different roles for those three collaborators.) You may also choose to work alone or in a group of two or four people. You may discuss this assignment and your work on the assignment with anyone, provided you credit such discussions when you submit the assignment.

Scope: You project should be of a scope that your group can complete it over a two-week period with approximately five hours of work per week per group member.

Introduction

At this point in the course, you have explored a variety of kinds of humanistic data and algorithmic techniques for dealing with them. We have, of necessity, examined only a limited range of algorithms and, even then, looked at only the basic issues associated with each. One of the more compelling aspects of the digital humanities, at least from my perspective, is that there is such a broad range of opportunities to do or discover something new. You can extend algorithms in different ways. You can tune algorithms to different kinds or collections of works. You can build tools that others may use.

Components

Part one: Project proposal

You will begin by forming a group, deciding upon an approach to the project, and identifying the materials that you will use as the core of your project. Please feel free to be creative in your choice of materials; while we have primarily been using works from Project Gutenberg and data from the US Government [6] in our weekly projects, you can use any collection that you can identify or create.

Based on those preliminary discussions, you will write a project proposal in which you provide a broad overview of the goals of the project, specify the materials you will work with, and describe the algorithm or algorithms you intend to implement.

That is, your project proposal should include at least three sections. A project statement, intended for a general audience, should explain the broad goals of the project. Why are you undertaking this study? What do you hope to learn? A materials statement should provide a more in-depth explanation of the materials you plan to use and where you obtained them. Finally, an analysis plan, intended for a technical audience, should describe the algorithms you expect to implement.

A typical project proposal will be approximately two double-spaced pages. You should write your proposal in narrative form, using correct grammar and spelling.

We will do our best to respond to your proposal promptly. However, given the time constraints of this project, you should move forward to subsequent steps before you receive feedback.

Part two: Project

You will have about one week to implement your project. That is, you will implement the algorithms you have designed, use those and other algorithms to explore your materials, and, one hopes, come up with some observations that can be used in a more comprehensive study.

You will submit your program code, the materials you analyzed, and instructions for using your program with those materials.

Part three: Report

A final project report should accompany your project code. This report will probably draw heavily on the project proposal. You may find that you need to modify your analysis plan to describe the algorithm you developed (plans and results often differ significantly), including details of the Racket implementation. In addition to the first three sections, you should also describe your initial results.

As you know, algorithmic analysis is only one step in many digital humanities projects. Hence, your report should also suggest subsequent work that might be done as part of a complete project. Have you discovered issues that would benefit from close reading? Do we need to look at other materials to learn more? Is it worth turning to the literature and placing this work in a broader context?

Part four: Lightning presentation

You will give a two-minute presentation to your classmates and a select group of visitors during the designated presentation days. We will also reserve approximately two minutes for questions and answers.

Selected approaches

Some of you will immediately identify a project of interest. However, others may need a few suggestions to get started. Here a few possible approaches you might find interesting. You will, of course, need to choose the materials with which you work.

Sentiment analysis. You might write an algorithm that takes a piece of text and identify the sentiment expressed in the piece. Is it positive, neutral, or negative? While much work in sentiment analysis is done using machine learning techniques (which we don’t expect you to undertake), you could experiment with algorithms that look at word choice and word proximity for certain words.

Topic modeling. You wrote a simple topic modeling algorithm for a homework assignment. You might extend that algorithm and apply it to a new data set.

Language generation. We’ve explored two models of language generation, one using generative grammars and one using statistical methods. You might look at ways to combine or extend those models, applying them to a particular genre or author. If you do choose a language generation project, make sure to reflect upon what we can learn about the source texts you’ve chosen; language generation that’s just for the fun of seeing what you get does not suffice as a project. It may be, for example, that you manually explore the structure of an appropriate sampling of sentences in the works and use those in your grammar.

Tool building. You might build a tool that other FunDHum students could use in future projects. If you choose to develop a new tool, you will still want to identify some materials that you can use to assess the efficacy of your tool.

Mapping. In our exploration of mapping, we read Barry Lopez’s The Mappist. Although you do not have the time to gather data of the kind that Benefideo used in his maps, you could, nonetheless, find ways of mapping things that reveal important and unexpected connections. If you choose this kind of project, one of your challenges will be to develop appropriately sophisticated algorithms; you already know how to map data, mapping alone does not suffice. You might, for example, develop algorithms that extract information from more complex data sources in preparation for mapping them.

If none of the above sounds interesting, you might spend some time exploring the Web for other possibilities. For example, HyperCities provides a wide variety of projects that you might find of interest.

We will spend some time in class discussing possible approaches to the project.

I also plan to put together a new rubric [7] and to write some extended examples of each kind of project. Those will likely have to wait until I write the rest of the course.

Now that I’ve finished this draft, I find that I also want to consider another kind of project. I’ll explain more tomorrow.

[1] And, on occasion, removed examples.

[2] More accurately, Janet or Jerod wrote a rubric.

[3] One of my Obermann colleagues has a variety of projects that explore language use in the Statistical Accounts.

[4] Of course, even though the range of possibilities seems limited for the MediaScheme project, the results vary widely.

[5] I’m glad to be done with way too many fractals.

[6] I need to spend more time thinking about the non-textual data that we will use in this course.

[7] I see that the rubric that Titus Klinge wrote for the data science version of 151 will be a pretty good starting point.

Version 1.0 released 2018-10-27.

Version 1.0.1 of 2018-10-28.