¶ 1 Leave a comment on paragraph 1 0 In this chapter, you have been introduced to a number of tools for fitting topic models to a body of text: ranging from one document, to an entire diary, to multiple documents. Different tools have contrasting algorithms that implement topic modelling in particular ways, leading to unique outputs. This is worth underlining! Indeed, this mathematical approach to human history does not produce the same specific pattern each time we run it, but rather a distribution of probabilistic outcomes. This can be difficult for us to get our heads around, but we must resist the temptation to run a topic model, and accept at face value its results as being ‘the’ answer about our corpora of materials. Turns out historians still have a job! Phew.
¶ 2 Leave a comment on paragraph 2 0 The topic model generates hypotheses, new perspectives, and new questions: not simple answers, let alone some sort of mythical “truth.” The act of visualization of the results too is as much art as it is science, introducing new layers of interpretation and engagement. On our companion website, we have written an essay that uses topic modeling to understand the historiographical discourses in the Dictionary of Canadian Biography, a multi-authored, multi-volume work that explores the lives of over 8000 historical Canadians (themacroscope.org/2.0/essays/8000.html). As you read that essay, consider for yourself the choices we have made in how we perform the topic model, and in how we visualize the results. How justified are we in those choices? How could different tools change our perspective and hence our conclusions? Do the visualizations help our argument, or hinder them? When you encounter someone else’s topic model, do not accept at first glance. Rather, to understand the potentials and pitfalls, you must be aware of how the tools work and their limitations. Andrew Goldstone and Ted Underwood published an article called ‘The Quiet Transformations of Literary Studies: What Thirteen Thousand Scholars Could Tell Us’, a topic modelling exercise on what scholars had written about literary history. In a companion blog post to that piece, Goldstone details the construction and implication of a ‘topic browser’ that he built to support the conclusions of the journal article. Goldstone writes,
¶ 3 Leave a comment on paragraph 3 0 [J]ust as social scientists and natural scientists grit their teeth and learn to program and produce visualizations when they need to in order to support their analysis, so too must those of us in the humanities. Solutions off the shelf are not great at answering expert research questions. What should come off the shelf are components that the researcher knows how to put together.
¶ 4 Leave a comment on paragraph 4 0 In the next chapter, we pause for a moment to consider ways all of this data could be visualized. The rhetorical impact of an excellent visualization should not be understated! This discussion of visualization and the principles of good design leads us to one particular kind of visualization whose popularity seems to be growing daily: the social network diagram. Social network analysis and its visualization have been conflated a great deal; yet it is entirely possible to have one without the other. We will argue in the final two chapters that historians need to understand social network analysis as a way of representing and querying the social relationships between historical actors in the past. And only then should we think about how to represent the network. As you will see, a ‘network diagram’ is often not the best choice!
¶ 5 Leave a comment on paragraph 5 0  Andrew Goldstone and Ted Underwood, “The Quiet Transformations of Literary Studies: What Thirteen Thousand Scholars Could Tell Us,” May 28, 2014, Preprint for New Literary History available at https://www.ideals.illinois.edu/handle/2142/49323.