I wonder how this rationale plays with the proposed goal of this work operating as a "counterhistory." The idea of this chapter (and others) seems to be to re-contextualize and add nuance and sometimes even problematize a visualization that is otherwise seen as "iconic." If you aren't already at least visually familiar with the image under discussion, does this chapter operate the same way? In short, is the content warning really a fair choice that will reduce harm, since the presumed reader has probably already experienced the chart under discussion? Or is it more a device to foreground the human suffering depicted in the visualization, with all of the spotty and mixed empirical records that other attempts to humanize data visualizations have had?

I'm having a bit of trouble with the argument here. It seems like, okay, the problem is that vis history is often laid out as this weirdly linear jump between "heroic" 19th-20th c. European examples (As I've said elsewhere think it's a lot like how the practice and history of medicine was viewed in medical history's own "heroic age"), and so ignores the rich history of representation practice outside of the Usual Suspects. But there could be several reasons for this:

1) Vis people aren't trained historians and so often suck at telling these stories. And so then the remedy is to just do what history as a discipline did when facing similar criticism about telling weird Eurocentric great-man-theory stories of history, which is to build stronger and more pluralistic methodologies. So then this book is about telling a "better" story than prior vis histories, but it's a difference in quality or methodology rather than subject.

2) No, vis really is special because "datafication" in a way that makes things legible to the practice of visualization is this situated and highly contingent process that is inextricably linked to a set of empirical and epistemic commitments and, furthermore, in a Marxist technological determinist way, to specific modes of production and social relation. In which case the project of the book is to make those linkages visible so we question the neutrality/teleological "progress" of data vis. So then the book is about a Zinn-style "counter history" of the existing story in vis histories. Which, fair enough, but then all of these examples that don't fit in the "conventional" story of datavis would seem to undermine the point. In other words, if "modern" conceptions of dativis are closely meshed with racist and colonialist and other specifically modern and western conceptions of information, why are there so many examples of things that look like modern datavis from outside of that specific context?

Probably not an answerable question, but what does restorative justice look like in this case? An analogous case of what to do with physical historical artifacts that are not "ours" (like the curation and display of First Nations artifacts) presents options of repatriation or reproduction. But what about the digital?

Ah, but that's part of the rhetorical goal of the chart, right? Not "here's a specific ship full of specific people" but "here is just one example, and think of how many other voyages are happening at the same time, with conditions just as bad or worse." I don't think this is a hill I want to die on, but I wonder if part of what makes the image so arresting (esp. to the intended [white] audience) is that the lack of detail allows one to expand from the specific to the general, and even fill in the details that are "missing" by design. Again, very thin rhetorical ice here, but you could imagine a sort of perverse increased ability to empathize by knowing as few of the details of the individuals as possible, and so being able to insert one's own particulars on the otherwise isotype-y figures, in a way one could do less well if each "data point" was imbued with a particular story or context that might be unfamiliar or far removed from the experiences of the proposed audience. I'm very uncomfortable with this point, hence all the hedging.

I think having an answer to this requires setting a moral frame, something which hitherto is implicit in the project of the book. E.g., some strawman utilitarian might say "well, looks like Diagram helped hasten the end of the Atlantic slave trade in the British Empire, and it didn't lead to any specific harm to any contemporary people at the time other than not individualizing them very well, which is a harm that is hard to quantify or even identify as having happened, so seems like the benefits here clearly outweighed the harms, next question." But presumably there's an assumption here that there are systematic complexities or omissions or harms that matter for certain moral frames that deserve to be foregrounded.

As mentioned in a comment above, I thought there would be much more on this point in this chapter!

One of the things I wonder, if we accept that data visualization inherently entails reduction and abstraction (and the task is therefore to have intentionality about what is reduced and why), if data visualization also inherently involves the distance of an observer and so the passivity that (can) come from that. You can chant a slogan as you march, or hold it on a sign (or graffiti it on a building). But you can't do that with a chart: you have to pause and interpret and decode it. It can be part of the rhetorical ecosystem that leads to political action, but maybe it's just a minor part. When I think of the what I retain from my US history classes on abolition and the slave trade I remember fiery rhetoric or harrowing individual stories; I don't remember data tables.

-> all

I think there's a bit more to say here about whether you thought this strategy was or wasn't successful for your goals. For instance, as mentioned, filtering only to voyages with resistance makes an implicit judgment: are voyages where enslaved people didn't resist (or resisted in less visible or datafied ways) of less worth? Or, even worse, the enslaved peoples in those voyages viewed as lesser for not resisting?

Or, as another example, as in Minard's map, deaths here are implicit: a narrowing and subtraction of each thread. If we can critique the description of a slave ship for abstracting away lives, what can we say of a visualization where deaths are not even directly visualized at all?

Other issues would be, as mentioned, that the intentional overplotting makes it difficult to find individual voyages, which might be specifically what I might want to do (for instance, to trace a specific path that one's ancestors might have taken).

I also wanted a little bit more about the non linearity of time here. It's just asserted as a fact (and I believe it, as per Drucker's discussions of time gets ordered and flattened in visualizations despite the myriad ways that time is experienced phenomenologically and cross-culturally), but I wanted more details for what non-linear aspects were salient here, and salient enough to intentionally foreground in the visualization, and important enough to "break" usual conventions of representing time. And whether this, too, was successful (to me, linear time is still very much foregrounded in the final design, in how one is meant to query and even "decode" the final strands).

Other options would be to, well, not visualize anything. To refuse to use these data just because they were available. Or to operate only with the participation, consent, or guidance of the descendants of enslaved people, and for specific rhetorical goals. "we would need a visual strategy" is actually several steps down the decision tree from those initial existential questions.

Gauche followup: how many of you had ancestors who were slavers

To me "this is agitprop with an explicit audience in mind" was front and center, so I was surprised to see the topic sentence of this paragraph. To me, speaking of the diagrams without grappling with "well, did they accomplish their rhetorical aim?" misses a key feature. Like, yes, of course they are reductionist and don't capture the individuality and interiority of their subjects. But, for their intended audience and intended rhetorical goal, did they need to? I think also worth considering how even other slave narratives (like Frederick Douglass') still implicitly or explicitly have rhetorical agitprop goals that also imply a focus on a (presumed white) audience. There's a sort of implicit line through this chapter that the design and datafication choices here were reductionist and wrong. But I wonder if the designers would have been fully cognizant of these choices and counter that the scale of the evil they were hoping to overcome was such that any means were justified if it would sway even one MP. In short, I wonder if the view that the voices of the enslaved are not adequately present in the visualization would be considered a minor sin, forgivable in the context of addressing the immense evil of the fact that the international slave trade was occurring at all.

I know DuBois had thoughts on this and other issues so I'll try to revisit this thought when I get to that chapter.

I'm interpreting the critique differently: that focusing the visualization on the flow of black bodies, the agency for who was responsible for enslaving and moving these people is lost. Less that the visualization gives false objectivity to what is situated knowledge, but more that the visualization falsely portrays the voyages as undirected and uncontrolled "flows" rather than intentional processes with specific malefactors attached. As an aside, I view many visualizations of climate data and greenhouse emissions as having this same category of deficiency.

Something fishy going on with the footnotes versus body text here.

This quote is reused above

double quotation marks?

Is this meant to be a footnote? I don't think it needs to be.

This quote was reused above.

The scrollytelling here wasn't working for me: was it meant to highlight portions of the document? I also feel like it has some redundancy with the text above.

Not grokking this distinction (both in that I don't understand what distinguishes an infographic from a data vis in general other than having different intended audiences and visual hallmarks, but also that I don't understand how a "direct representation of data" is one of those hallmarks).

I agree, but I do wonder something. Symphonies, paintings, poems, etc. all can support a richness of sensory and emotional interpretation, beyond even what the designer may have intended. It seems that we can "add nuance back in" to these kinds of artistic forms by taking time to situate the text within contexts, add our personal histories and emotions to them, etc. But data vis is purported to not support this kind of enrichment. I wonder why? Is it "datafication's" fault?

I think there's a way to make this point without bringing Tufte into this chapter. I don't need him here.

I don't quite buy the gloss around it, but the isotype examples from the Soviet Five Year plan I think are an easy sell for "using the abstractions afforded by isotypes for political projects" if you don't want to use the same cartoonish racist caricature isotypes that other people seem to use. https://medium.com/nightingale/exploring-soviet-isotypes-digitizing-the-struggle-for-five-years-in-four-50df7417a766

Thinking here of Kennedy, Helen, et al. "The work that visualisation conventions do." Information, Communication & Society 19.6 (2016): 715-735. for more on how clean lines and boundaries in data vis perform rhetorical work.

Thinking of Bloom Theory again!

You reuse this quote earlier.

Despite having the content warnings turned off, the scrollytelling for this section is still focusing me on blurry parts of the diagram; it's only when I click the image that I get the actual full resolution diagram.

I don't think I like the partial reveal scrollytelling approach to the diagram. For one, (as RJ mentioned in the past chapter when a similar trick was done to Minard), it sort of cheapens the work to think that a modern viewer would need some additional SVG elements to highlight. For another, it dampens the immediate, arresting nature of the graphic that both the chapter (and the original creators) wished to instil: of the mass of human beings crammed like cargo. Showing just one "section" at a time lessens this impression! If kept, I'd prefer full diagram first, then scrollytelling to highlight and discuss individual portions after.

I felt a tension in this chapter. The intro seems to be hinting that we need to build a more holistic view of data vis history beyond just the same old examples and acknowledge the other voices and contributions and implicit sociopolitical underpinnings that enable particular kinds of charts and graphs. But this chapter didn't feel "counter-" enough to me. The status quo is that Playfair was an important guy who invented a bunch of the standard genres of visualization but was not heralded in his own time, and this chapter seems to mostly reinforce that narrative (even though I'd challenge it, especially with things like bar charts where there are several pre-Playfair examples).

Same note as in earlier chapters; I'd prefer the "original" chart(s) first, and then the SVG decompositions after.

I'll got a step farther and say that it's also about the patterns of social relation that those tools and technologies support or reinforce.

Am I missing something or did we not get told what the actual error was yet?

Not just objectivity but also, as per Shannon Mattern's work, an explicit and externalized ontology. For instance, the racial categories used in census data around the world are inherently political projects with explicit repercussions in how they shape the nations in which they are used. Spreadsheets also implicitly create blank spaces to be "filled" (in similar ways that maps create "here they be dragons" unknown regions that must be explored). This interview from Sarah T. Hamid: https://logicmag.io/care/community-defense-sarah-t-hamid-on-abolishing-carceral-technologies/ has this quote that I keep reusing: "There were blank cells in the spreadsheet, and we became obsessed with filling them in. And then after a week we were like, 'Why are we doing this? Why are we so obsessed with having a complete spreadsheet?' We started to realize that our way of knowing and our mode of inquiry were being influenced by the nature of the spreadsheet. It wasn’t curiosity, or any real need to find the information. It was the structure of the technology. Knowledge takes a particular shape when you start to use particular mediums."

I'm not quite sold that this is a specifically colonialist and post-renaissance view of data and dominion. I mean, the notion of a statistical census of the population recurs in pretty much every imperial power from Mesopotamia to China to Rome to you name it. Keeping track of debts and land and so on seems to happen across large swathes of societal modes, in other words. I think there is something to say about how some of these forms are more or less readily visualizable, however. And example we encountered when looking at how people across cultures visualized disease data were the Ming dynasty "gazetteers." Tons of data collection in service of an imperial project but, as per Dunstan, Helen. "The late Ming epidemics: A preliminary survey." Ch'ing-shih wen-t'i 3.3 (1975): 1-59., a total pain to turn into something like a double column table or a thematic map, because the structures of the gazettes were formalized and included mixtures of what we'd consider "reporting" with things like poetry or literature, without clear ways to aggregate these gazettes.

Obligatory Tufte hot take(s): one of the things I dislike about Tufte is that the style of argumentation (often very useful examples supported by argumentation that is both often very dogmatic/bombastic but not necessarily strongly empirical, is I guess the politic way I would put it) is that it immediately prompts a strong desire for an antithesis. "No Tufte, it's not so simple as that!" Or "No Tufte, chartjunk is sometimes beneficial." But I've become uneasy with either Tufte-boosting or Tufte-bashing.

For one, whenever I see thesis/antithesis the residual Marxist/Hegelian in me wants to explore what the synthesis looks like. E.g., rather than "chartjunk is bad!" "Chartjunk is good!" I think I've come around to something like Akbaba et al's thesis in their manifesto that chartjunk is a not a conceptually useful category and should be excised from academic language entirely and replaced with a more useful debate, around, say, when and where to employ complexity and adornment in information design. It's just more productive, to me.

For another, I wonder if there's some weird aspect of situated learning to Tufte's work. Like, all the weird non-empirical stuff that results in howlers like the error bars or box plots with random lines removed as allegedly "good" examples of redesigns to reduce data-ink are perhaps not compelling, but if you're, say, somebody in big business or government and the default mode of communication is a series of interminable powerpoint decks with logs and animations and gradients everywhere, maybe that sort of reductionist minimalist style is actually what you need to hear (with, again, the assumption that presumably you'll arrive at some synthesis later).

Love this point: I really hate when the history of visualization is reduced down to a single linear teleological intellectual project, where early visualizations are seen as naïve, and there is no path forward other than increasing technological complexity, data collection, digitization, etc. I'm reminded a bit of the old BBC connections series with James Burke: showing how often chance links and connections across time create a multipolar web of scientific achievement rather than a linear directly causal narrative.

And there's Chinese Counting Boards, and Marshall island stick charts, Nile-o-meters, and maps of course that in some cases predate writing...

Although there is an interesting question (and I think the Wainer/Friendly book gets at this) about the distinction between representing or exteriorizing information (for instance, are ideograms or writing systems data visualization?) and the vicissitudes that the definition of "data visualization" has to be put through for claims like "Playfair invented the bar chart" to be close to true. So stuff like Philippe Buache and Guillaume de L’Isle's 1770 chart of water levels in the Seine is said not to count for the "Playfair was first" bar chart narrative because the bars represent some physical quantity (river height) rather than an abstract quantity (Playfair's exports and imports). Likewise I wonder what other definitional challenges exist mostly to exclude other ways of externalizing information. Or things like Snow's Cholera presented as this notable first for data visualization when similar dot maps of epidemiology data were created (e.g. Valentine Seaman's yellow fever map) decades earlier, but don't get credit mostly, it seems, because the conclusion of that map (that miasma caused yellow fever) was /wrong/ whereas Snow just so happened to be /right/?