I love to cook. My wife says I’m pretty good at it, although that could just be an attempt to not have to cook herself. We don’t have cable so I don’t get to watch many cooking shows. However, occasionally when we travel, I get the Food Network… then I become a couch potato. I’m hooked on watching people compete by preparing different dishes in different ways. I especially love the ‘random box of stuff’ sort of competition where the chef needs to work out something delicious from a collection of ingredients. That’s some real creativity there.
What I don’t love about these shows is the judging part. This collection of experts sits down and looks at their food, smells their food, and tastes their food rendering a ‘this is better than that’ decision. Now I’m no expert (I’m hardly a decent cook, much less a chef or a food expert), but it all seems so arbitrary to me. However, I’ve noticed a bit of linguistic turn coming into the descriptions from the judges. There are what I’d dub ‘axis’ words, such as ‘acidic’, ‘sweet’, ‘sharp’, and ‘bright’. Then there are more descriptive words that seem to modify that a bit, such as ‘tangy’, ‘flavorful’, ‘steep’, ‘heavy’, ‘light’, etc. Finally, there are some comparison words used, most normally ‘balance’ and ‘counter’. So you might hear a description such as, “I like that you balanced the heavy sweetness with a flavorful acidic quality.” To my quantitative ear, it almost sounds like these professional chefs and judges have this sort of equation in their head that moves along the axis of acidic, bright, and sweet. It is as if there is an attempt at an extremely loose quantification of what is basically a qualitative thing. How do you impose a sort of equation on what’s basically opinion? I don’t like raw tomatoes, so anything with raw tomatoes is going to trend ‘negative’ for me personally despite their utility in providing an acidic quality to the food equation. The point being, any one person’s mileage may vary here.
Now let’s jump to geography and GIS. Lately I’ve been straddling the divide between qualitative and quantitative data. We like quantitative data because you can measure it, you can compare it, you can mathematically transform it, if it relates to space you can map it, you can color it…. you can do all sorts of things to it. Not only that, we can represent that in known ways with agreed upon conventions. Things like gradients or relative shape sizes have been reasonably well worked out. We tend to know what to expect, and in fact we recoil when they’re not what we expect. Quantitative data doesn’t present us with much challenges. Qualitative data is a whole ‘nother critter. We don’t always know what to do with qualitative data. We can’t even agree if it has much utility or not (editors note: WHAT? Blasphemy!). We certainly don’t know how to store it in transformable forms, or even how we’d like to transform the information. We don’t know how to compare it, or even if it is comparable. And the issue of representation? That’s so far out there and varied it’s barely on the radar.
You can certainly see why judges seem to be attempting to ‘quantify’ these qualitative measures. With numbers you can work out a ‘winner’ – feelings, emotions, tastes are a LOT harder because of their subjectivity. However, I think we lose so much when attempt to boil down these complex flavors and aromas into a comparable framework. How can we capture that information and still retain the ability to compare, contrast, and represent in agreed upon ways? I’m going to do the thing I hate the most and cop out because I don’t have an answer. To be honest, I don’t think there is ‘an’ answer but a series of answers we have to work out.
One of the things I love most about Exploratory Data Analysis is that at its core philosophy it says, ‘don’t mess with the data’. It is about exploring the data, the connections contained therein, the trends and implications, and the information all that embodies, but doing so in a way that preserves the data. So we can transform, reproject, represent and re-represent, and move around the data, but we try not to change its fundamental character. For me it’s like moving furniture around a room – the layout of the room changes, but the basic function stays the same. We need to implement the same philosophy when dealing with qualitative data. The information stays the same but we should be able to move, alter, transform, shuffle it around, and all the sorts of things we like to do with quantitative data. How do we allow both the ability to ‘push the furniture around’ and transform the data like we normally do with quantitative data?
Most people can see that 1 is the same as 2/2 and that’s the same as 173/173. We’re comfortable with those sorts of re-representations on the fly. We’re also pretty comfortable with changing the colors on a map as long as it is reflected in the Legend. If the blue dots become red dots, who cares? This sort of re-representation and re-visioning of quantitative data is pretty standard stuff. But how do we have multiple forms of representation with qualitative data without fundamentally changing the nature of the data? Different people get different bits of information from qualitative data because they just see different things in it. It’s important to be able to spin it around, twist it, and push it about until you get that ‘a-HA’ moment yet retain everyone else’s ability to do the same. Saying, ‘That’s tricky.’ might be the mother of all understatements. How do you take some text and make it a picture, or vice versa? Could you even begin to do that? Do you even need to do that? How can you cross link to information that’s related to the same thought space without having to know the sum of everything that’s out there? Now we’re dipping into the area of semantics, phenomenology, subjectivity, positionality and a whole other can of worms.
Like I said a couple paragraphs back – I don’t have an answer for all this stuff. If I did, I’d be a bazillionaire 🙂 However, I think we have to look pretty critically at attempts to ‘equationize’, if I might invent a word there, qualitative data. Yes, we can say that most people would find a master chef like Gordon Ramsey’s food more enjoyable, more tasteful, more artful than my own, so maybe we can compare two qualitative values. However, we’re talking about extremes here. Seems to me that the devil is in nuances, but so is the really good stuff. All I know is qualitative data is going to be critical in the future and people much smarter than me need to be attacking these issues head on. Luckily, others have started working out some of this and I’m excited to see what they come up with in the future!