How was the last exercise? Any attempts?
Here’s the gist: we took four days of data on our sensory experiences during meals. Although we tend to associate food and eating under the umbrella sense of taste as an experience, each day we focused on collecting data from the different senses we encounter while eating:
Phase 1: Data collection
Day 1: texture
Day 2: color
Day 3: sound
Day 4: taste/smell
The end of day four was bittersweet: the exercise was simultaneously fun and exhausting. I was relieved to prioritize acquiescing my hunger and not scrutinizing my sensory perceptions. Before we could move on to the next sensory challenge, we decided to analyze and document our experience through a 4-step process developed by Jordan.
Phase 2 : Data analysis
Sound or gesture data representation
I’ll go through these one-by-one combining examples from our own accounts in order to illustrate how diverse the expression of this data collection exercise can be. Here in part 1, I chose an example from each of our data analysis exercises from Phase 2 and paired them with data collected from one of the senses during Phase 1. If that sounds confusing, here is a convenient wayfinder for this article that highlights the data (textures and sound) and analysis (reflection and data entry) that we will explore during this issue. The next issue will touch on the scroll, which integrates all of the days of data collection, and the gesture/sound that is exemplified using taste.
Prompt: Reflect on the experience of collecting data about taste and write a narrative account of what it was like. Don’t worry about the quality of the writing at this point. Think about this more as a free-writing, stream-of-consciousness exercise. Tell the story of the data collection.
This one felt really difficult, due to the lack of vocabulary for textures. I found myself using the same texture words to describe different foods in my mouth, and although I knew the nuances existed, I couldn’t quite place them. It made me think of Japanese onomatopoeia, of which there are at least a thousand that are used in everyday life. At the same time, by doing this exercise, I became more aware of which texture categories appeared most frequently and my frustration made me search for other, more nuanced words. As a result, some of my textures are more specific than others (to avoid repetition), such as with “crunchy.” (Excerpt from my Day 1 reflection)
Characteristically, my mind turned to a Japanese anecdote, specifically Japanese’s onomatopoeia, and the time when I asked my Japanese friend about the translation for ‘crunchy.’ “Which one?” She replied. Indeed, in Japanese a crunchy cookie is zakuzaku, while potato chips might be described as baribari, rice crackers as paripari, apples as karikari, boribori is munching on something harder, like ice, and zakuzaku is walking on something crunchy, but I was actually looking for shakishaki, which describes that satisfying crunch you get from biting down on a fresh cucumber. Shakishaki. We know that data collection is imperfect, and the lack of appropriate and nuanced vocabulary is a testament to it. Vocabulary, whether precise or imprecise, adds its own biases onto the data and the data that could never be.
While I focused on the textures I encountered while eating, Jordan considered even more philosophical questions.
As I attempted to write down the textures, I realized how complicated this seemingly simple task was. There wasn’t just one texture, but many. There was the texture that I could observe, with my eyes, as well as the texture that I could feel with my mouth, lips, and tongue. Plus…how does temperature fit in? Is heat an element of texture? I tried to push all of these questions aside for now and just jot down what seemed reasonable: (luke) warm, wet, liquid, fluffy/frother on top, then below shiny/reflective, chunky (although I don’t feel the chunks because they dissolve in my mouth). These were the texts that I felt with the nerves in my mouth. Then for visually observable textures, I wrote: (mottled?) and tried to sketch what I saw – which looks kind of Dr. Seussy, a squiggly line with little tufted balls breaking it up. (Excerpt from Jordan’s Day 1 reflection)
The reflection activity was an appropriate start and turned into a foundation we could build on.
Max: After writing this [reflection], it didn’t feel complete. I ended up adding to my reflection while doing the subsequent exercises just because the act of exploring different expressions of data gives you a new perspective and opportunity for reflection.
Jordan: Like turning it into a table - I came back to the reflection because the table jogged certain parts of my memory. The act of doing this first…
Max: …set the stage for the rest of the exercises. [Excerpt from our post-activity conversation]
Reflection, whether captured in writing or by voice immediately after data collection should be woven into the process. I wouldn’t say this stage is where most of the insights emerge, but it is where we noticed them take on a cohesive form out of unorganized idea threads and provide a space for expansion as we continued to work with the data. Yet, it also is a foundation for further analysis, data representation, and, of course, reiterations for the data collection exercise itself.
Prompt: Take all of your data collection notes and put them into a spreadsheet. You get to decide what the data structure may be (i.e., what the columns and rows are) and what variables should be included. If you try to put your data from all of the various days into one schema, this may require you to create new, interpretive variables in addition to what you collected differently, or to code your qualitative data in a new way.
In this excerpt from our conversation, we grappled with the decision about what to include in our structured data tables:
Max: I knew that for however many [metrics] that I’d write down, I wasn’t going to be able to capture the experience and so that was just like for me incredibly frustrating.
Jordan: You have to pick within this infinite possibility to focus. Then there’s that kind of correction like “Oh, am I focusing on the right things?” and because we left it sort of open. There’s so many different dimensions of the sensory experience and so there’s so many different things that you could collect data about and then so many different ways to represent it. It’s both exciting, but then also, I think, a little intimidating. How do we make sure that we’re doing this in a way that is, I don’t know, if it is authentic to the sensory experience?
Once I accepted the pessimism and imperfection, I decided to split each day of data collection into its own spreadsheet tab, customizing the metrics to suit my imagination and my attempt to describe the experiences. While all tabs include a description of the meal, they diverge thereafter. Some metrics I created during this exercise, some I developed for later exercises, such as in the 4th where I would create a gesture/sound out of the data. Consequently, my data wasn’t as structured or fixed, as data is normally thought of. The metrics mutated as I logged them and as my experience faded into retrospection: they were fluid, constantly being iterated upon to fit a purpose.
Below you’ll find an excerpt of the sound tab on one of my datasheets. The different metrics I added (origin of sound, touching instead of hearing) were included after the fact as I prepared for the following data representation exercises.
What kinds of metrics can you take from sound? I approached this question inversely: I thought about sonification, translating data to sound, whereas in this exercise I was translating sound to data. Sonification is often composed of data over time: sound is often thought of as being linear with a clear beginning and end. This is perfect for timelines where the listener is guided on a musical path that could be visualized as an oscillating line across graph paper. Nevertheless, I’m attracted by the idea that sound is spatial, and so I decided to play with encoding sound beyond its linear composition. For ‘origin of sound,’ I noted the source of the sounds. “Internal” meant that I made it while eating; it came from my mouth. “Immediate vicinity” was more fluid, primarily external sounds that I caused, such as a spoon clinking against the bowl.
For the column, “touching instead of hearing,” I tried to think about how I could represent this data as an interactive experience. Ultimately, I imagined that people would touch these objects, and it would remind them of the sound that I experienced.
Jordan, however, chose a different route: she collected data on the food item, meal, if she recorded it retrospectively (after eating), where she ate it, and to what extent she felt she was able to capture the experience.
Jordan’s only non-categorical entry was, “To what extent did I fully capture the experience,” which was an ordinal scale of 1-10. With this metric, she grappled with how all data we collect is a representation, or a translation, of lived experience. The best score she gave herself was an 8 out of 10—conveying how there will always be a gap between an experience and data about that experience. For most of the entries, she also consciously chose not to write down what the food was to see if the context around it was enough to jog her memory. This challenge used intentional sparseness to interrogate the gap between the data collected, her lived experience, and her memory of the experience.
The fact that my spatial data collection is more granular than Jordan’s general location metric reflects the importance of how contextual data is, and what goal we are trying to achieve with it. Jordan sought to illustrate her movement while eating, though I always eat at home during the week when I’m in California, and an entire collection of “home” labels is arguably far less engaging and insightful than data of the sounds’ spatiality. I was forced to notice which meals tend to have which sounds: the mornings are the most tumultuous with dogs, coffee makers, etc., while my lunches outside are [more often than I would like] characterized by one or two ferocious lawnmowers. I often take my desserts in my room, which is why more “internal,” and therefore intimate, sounds are present. Documenting the origin of sound was like painting my daily mealtime soundscape.
This data entry exercise and the reflection exercise set the groundwork for the scroll and sonification exercises, which represent that data in sensory forms. Until we reveal these results in the next issue, I’ll share some insights on what we learned from collecting the data. Challenges of data collection
As data nerds and multisensory enthusiasts, we were eager to begin this data collection journey. Nevertheless, even we found that this task was challenging and were forced to accept the imperfection of the data. Since we started the data analysis exercise a week after our data collection phase, we relied on memory to retrospectively add metrics and data, as shown in this excerpt from our post-exercise conversation:
Jordan: I thought that I would do a better job when I was at home versus when I was at work but that actually didn’t really end up being the case. It had to do more with what I was collecting, like whether I was focusing on the sound (easier) or the texture (harder).
Max: Yea, I ended up filling things in retrospectively too, or omitting my snacks. Those were more spontaneous and I just, you know, I just wanted to be fast. I was hungry, I just wanted to eat something; I was like no I can’t do this! There’s only one that I took down because I felt like it was really interesting and I wanted to document it. [Embarrassed laugh]
We both struggled with how much we could have taken down, worried we were omitting something that might be important, while knowing that having too many metrics would drown out the others, diminishing their importance. We juggled with using filters effectively to be simultaneously thorough and concise.
Jordan: Data collection is always a filter so what kind of filter are you doing and how are you choosing it? I had to let a lot of it go and accept that it is what is…there is no perfect and no complete.
Max: Amen to that.
What did you or would you do? Separate data sets? Paste them together? What metrics would be useful for you to write down?
Stay tuned for part 2 of this exercise: scrolls, sounds, and summaries.