This title might seem to mention two different subjects but they do naturally link.  

I recently posted about a tea conference event in Singapore.  That was based on input from two acquaintances who attended, where the organizers (Teapasar) used an aspects range profiling system to help guide participants in what to try.  The input was from an online survey of preferences, about food types and such, which was converted to a profile mapping and set of tea recommendations. Very cool!

Teapasar profile system summary

This reminded me of getting into the subject of machine / sensor testing of tea flavor aspects last year.  That started more from the starting point of potentially measuring tea quality, more through markers than detailed analysis, but the review scope gradually shifted into considering general tasting.  Can a machine (sensors, and AI / artificial intelligence programming) “taste” tea? Not really, but there is research on a starting point, that gets further than one might expect. Explanation works better based on background about what people perceive when they taste foods first, but this is just a broad summary of some related ideas, not including that part (covered more in the second of two original posts on aspect profiling and machine tasting).

Tea aspect profile mapping

The Singapore event Teapasar organizer describes the system in general terms:

Our ProfilePrint methodology also identifies distinct taste profiles of each tea listed on the marketplace. At the same time, when customers create their personalised taste profile online, their unique preferences can then be matched to our database of teas, and the closest matches can be recommended.

Cedric Teng related his impression of it at that Singapore event:

…users were prompted to select their tea preferences based on 8 different tea characteristics, namely sweetness, umami, saltiness, sourness, astringency, bitterness, aftertaste, and richness. The online algorithm will then formulate an individualized Profile Print based on these preferences, and offer some suggestions for teas that you may enjoy and the percentage match…  With the number of booths available, it soon became more convenient to just go from booth to booth rather than follow the website recommendations, and I found myself sampling a bit of everything…

…Even though I started off preferring teas which were stronger and more bitter, I found myself gravitating towards sweeter teas with floral or fruity notes in the end, but I think more tea needs to be drunk before I decide on my preferred tastes.

Seemingly a well-organized event

It would be natural to want to try as much as possible of samples available at booths, versus working around any preconceptions or aspect mapping formula.  It’s a much different context than ordering tea samples online.

This is actually the second such tea aspect profiling system I’ve encountered, the first being early development of flavor mapping conducted in a Penn State Tea Institute group (my alma mater, by the way; I graduated in Industrial Engineering there long ago).  That earlier version was an tea tasting oriented approach that later initiated a private food-service profiling system, the Gastrograph application (available through Google Play), which does essentially the same thing the Teapasar ProfilePrint system is doing, just not limited for use for tea.

A Guardian article reviews this general theme and specific program features:

…Analytical Flavor Systems’ main data collection tool is its smartphone app, Gastrograph. The app’s central feature is a wheel with 24 spokes, where each sliver represents a discrete category of sensory experience – such as “meaty”, “bitter” or “mouthfeel”. Tasters map the contours of flavor perception by tracing the spokes corresponding to the qualities they detect, designating the intensity of each on a scale from one to five. A submenu allows for a more granular record of experience…

I did check out that app.  It’s simple to use, just not functional without more purpose or context.  There would need to be some equivalent to using a database to get suggestions, as in the Teapasar case.  If it was more customizable in format and export function it might work to retain tea reviews but written notes would seem more functional than a limited graphical list of aspects.

To be concluded tomorrow