Monday, July 22, 2019

Perfume and Big Data: Garbage In, Garbage Out



 A new “big data” study of perfume has been making a splash the past couple of weeks. Titled “Social success of perfumes,” it was published in PLoS ONE on July 4. It’s another in the recent genre of throwing abstruse mathematics at the sense of smell.

The authors, Vaiva Vasiliauskaite and Tim S. Evans, are in the physics department at Imperial College London, more specifically the Theoretical Physics Group and Centre for Complexity Science. [Awesome nameage!—Ed.] Vasiliauskaite is a graduate student and apparently a talented nerd, having graduated from the University of Glasgow with a degree in Theoretical Physics. Evans is a Senior Lecturer.

Here’s the abstract of their paper:
We study data on perfumes and their odour descriptors—notes—to understand how note compositions, called accords, influence successful fragrance formulas. We obtain accords which tend to be present in perfumes that receive significantly more customer ratings. Our findings show that the most popular notes and the most over-represented accords are different to those that have the strongest effect to the perfume ratings. We also used network centrality to understand which notes have the highest potential to enhance note compositions. We find that large degree notes, such as musk and vanilla as well as generically-named notes, e.g. floral notes, are amongst the notes that enhance accords the most. This work presents a framework which would be a timely tool for perfumers to explore a multidimensional space of scent compositions.
Leaving aside the technical terms “network centrality” and “large degree notes,” the claims in the abstract seem clear: the authors have identified odor descriptors that drive the commercial success of specific perfume formulations. Sounds interesting and useful. One dives eagerly into the paper to find the details.

That’s where the problems start.

No actual perfumes were smelled in the making of this study. Nor were any actual perfume formulations examined. Instead, the authors apparently scraped perfume description and ranking data from a website, cleaned it up a bit, and then proceeded to slice, dice, and theorize.

I say “apparently” because nowhere in the paper do they describe where or how they obtained their data. I imagine my amateur readers stammering, “B-b-b-but, don’t scientists have to describe their data?” Yeah, well, uh, no, I guess not.

From clues in the text (e.g., perfumes rated on a five-point scale), Vasiliauskaite and Evans may have tapped Fragrantica.com as their source. In which case, they might have been courteous enough to give the site’s proprietors a shout-out. Or, better yet, have obtained permission to use the site’s data.

Of greater ethical concern is the absence of the source data in a supplemental file or online scientific archive. Making the data available is a requirement of publication in PLoS ONE.
PLOS journals require authors to make all data underlying the findings described in their manuscript fully available without restriction at the time of publication.
I don’t understand how this basic requirement could have been overlooked by the reviewers or the paper’s editor, Yongli Li of the Harbin Institute of Technology in China.

So, let’s look again at the abstract. By “successful fragrance formulas” the authors mean perfume brands with a large number of user reviews and high ratings on an undisclosed fragrance website. By “notes” and “odor descriptors” they mean descriptors provided by anonymous reviewers of undetermined skill level and/or marketing verbiage lifted from advertisements and promotional copy. On Fragrantia.com, gauzy words like “honey,” “amber” and “musk”—which refer to no one specific perfumery material—are given equal weight with clary sage, tonka bean, and oakmoss (see, for example, Paco Rabanne Pour Homme). By “accords” the authors appear to mean co-occurring “notes” on a webpage.

Other passages make one wonder if the authors know what they’re talking about. Take this paragraph, for example:
Information of the precise amounts of each ingredient in the formulation of a perfume is confidential, to prevent duplications of the formula. However, the list of ingredients, the list of notes, is often advertised in order to describe the scent of a perfume. Thus a perfume which smells of rose, vanilla and musk, is described using such notes. In this study we have analysed the notes which make up over ten thousand perfumes without knowing anything about their specific amounts in each perfume. We assume that a note is included in the perfume description as its presence enriches the composition and its smell is detectable.
This is paragraph is confused to the point of idiocy. A perfume’s list of ingredients is never revealed, much less advertised. What is publicly promoted by the brand is a short list of note names meant to imply romance, exotic origins, and high quality as much as what the perfume smells like. Even then, it’s unclear whether the “notes” used in this study were provided by the brand or plucked from the website’s crowd-sourced reviews. The authors “have analysed the notes which make up over ten thousand perfumes”? Fairer to say they’ve analyzed notes attributed to perfumes (by persons unknown) that may or may not include all the salient notes in a given perfume. The stated assumption that each note in a description is individually “detectable” is ridiculous. Do the authors believe that the grapefruit and Calabrian bergamot in Coco Noir are individually detectable? One wonders how much smelling they’ve ever done.

At this point it’s clear that no matter how much “network” and “non-network” statistical analysis they apply to this slop bucket of data, the results will lack specificity and insight. For all of its nodes, edges, weighted network representations, permutation tests, d-scores, and one-mode projections, the study doesn’t make much contact with the commercial or sensory realities of perfumery. It could have been interesting as an analysis of brand attributes seen through social media. But even there it fails.

This paper isn’t worth the time it takes to download.


The study discussed here is “Social success of perfumes,” by Vaiva Vasiliauskaite and Tim S. Evans, PLoS ONE 14(7): e0218664.

Sunday, May 12, 2019

Did Nicolas Guéguen Really Replicate the Cinnabon Effect?



Some years ago I blogged about a couple of scent-related field studies conducted by French academic researcher Nicolas Guéguen. In 2009, I wrote about a paper of his I had just come across, published in Psychological Reports back in 2001. In “Effect of a perfume on prosocial behavior of pedestrians,” Guéguen found that people were more helpful to a young woman who “accidently” dropped an item when she was wearing perfume, compared to when she was not wearing scent. I thought the paper was worth mentioning because it described a theory-free, empirical approach that could shed light on the natural history of perfume use.

Three years later, in a post called “The Cinnabon Effect Confirmed,” I wrote about another of Guéguen’s studies. This one, published in the Journal of Social Psychology, was titled “The sweet smell of . . . implicit helping: effects of pleasant ambient fragrance on spontaneous help in shopping malls.” The original Cinnabon Effect study was discovered in 1997 by social psychologist Robert Baron who found that shoppers in a mall were more helpful to a stranger when approached next to a pleasant-smelling retail shop (e.g., Cinnabon) rather than an unscented one (e.g., Banana Republic). In his new paper, Guéguen found similar results in a French mall, and extended the paradigm to include male as well as female “strangers.” I wrote approvingly that “Guéguen has confirmed the Cinnabon Effect and extended it to spontaneous acts of kindness.”

I recently became aware that Guéguen’s “sweet smell of” study has been the object of some impressive data sleuthing by Nicholas Brown and James Heathers. A new post on the excellent Retraction Watch website updates the story. In a nutshell, the pair claims that a bunch of papers by Guéguen don’t pass the smell test. The reasons for their skepticism are procedural and statistical. Guéguen is a prolific author yet one with no collaborators, and he may, according to Brown and Heathers, have skirted ethical requirements. Disturbingly, statistical patterns in his underlying data seem consistently fishy. The original B&H posts can be found here and here.

It’s important to note that Brown and Heathers do not claim outright that Guéguen’s studies are fraudulent.
We have not made, and do not make, any specific allegations of fraud, nor are any implied. The initial document that we released is entitled “A commentary on some articles by Dr. Nicolas Guéguen” and details a long series of inconsistencies in research methods, procedures, and data. The words “fraud” and “misconduct” do not appear in this document, nor in any of our communications with the people who helped with the investigation. We restrict ourselves to pointing out that results are “implausible” (p. 2) or that scenarios are “unlikely [to] be enacted practice” (p. 31).
You can download their detailed fifty-two page critique here and judge it for yourself. If you’re less hardcore, this Ars Technica piece covers the story nicely.

I think the points raised by Brown and Heathers are compelling and cast serious doubt on some, if not all, of Guéguen’s work. Accordingly, I’m putting a metaphorical asterisk on the two studies I cited, and flagging this action on my previous blog posts. Data integrity and transparency are core to the scientific endeavor, and when serious doubts are raised they need to be acknowledged.

Thursday, January 17, 2019

FN Review of The Coffee Visionary: The Life & Legacy of Alfred Peet



On my desk as I write this is a mug of Peet’s Costa Rican coffee, ground in my kitchen, brewed strong, and drunk black. The beans arrive via UPS within days of roasting. Peet’s has been my house coffee since I was a Cal Berkeley undergrad. I remember seeing Alfred Peet wearing his shop coat behind the counter at the original store on Vine and Walnut. So, in the interests of full disclosure: yes, I’m a proud Peetnik. Peetniks are all about the coffee: the beans, the roast, the brewing. Like Alfred Peet himself, they disdain what they regard as the coffee-flavored milk drinks served by Starbucks.

Despite my lifelong commitment to the brand, it turns out there was a lot I didn’t know about Alfred Peet. That’s been remedied by Jasper Houtman’s excellent new biography, which covers Peet’s personal life as well as the history of the company. Houtman bases his account on published and unpublished interviews with Peet, as well as first-hand interviews with many people who knew the man personally.

Born in Holland in 1920, Peet was raised by a stiff, demanding father with whom he never established a warm relationship. Years later, when he traveled and lived around the world, he wrote only to his mother, never to his father. This alienation wasn’t entirely his father’s fault. Young Alfred was a rebellious child: he didn’t do well at school because he felt pressured to perform, and he chafed at the tight confines of bourgeois Dutch culture.

Peet’s father and various in-laws were in the coffee business in Holland. Alfred had a natural flair for tasting coffee and tea, but rather than join his father’s company, he left Holland as soon as he could. After WWII went to London to work in tea for the Lipton company. Soon he left for the Dutch East Indies (today’s Indonesia), and after that for New Zealand. Seeking a still bigger horizon, he came to the United States in 1955, landing in San Francisco. A series of short-term jobs—including door-to-door salesman—followed, which the shy Peet found excruciating. But he was determined to make his way.

At the time, most coffee came pre-ground in cans from the grocery store. But having discovered the Italian espresso culture of North Beach, Peet realized there was an opportunity in America for better coffee. He was determined to make it happen. Unlike the Bay Area innovators of recent years, Peet was no young upstart. He didn’t open his first store until 1966, when he was forty-six years old. His timing and choice of location couldn’t have been better: the store on the corner of Vine and Walnut became a founding feature of Berkeley’s Gourmet Ghetto. The area’s reputation was cemented when Alice Waters opened Chez Panisse around the block in 1971.

From the beginning, Peet’s mission was to teach people about coffee. In doing so he didn’t pull any punches. He railed against customers who brewed his coffee in percolators. He fired employees who didn’t follow his strict procedures. And yet he freely shared his knowledge with people who wanted to learn about coffee. He trained many coffee roasters who went on to build their own businesses (such as Starbucks). Despite his Dutch directness—which Americans often read as rudeness—and his inflexibility about the correct ways to do coffee, many of those he trained came to fondly regard him as a mentor.

Like most entrepreneurs, Peet assumed total control of all details of the operation. Part of this was intellectual—the man was a perfectionist who insisted on doing it the best way, or not at all. Another part of it was characterological: scattered throughout Houtman’s book are examples of obsessive-compulsive behavior, as well as a compromised approach to interpersonal relationships. These traits enabled him to create a brand known for its commitment to quality, but also limited its potential to expand. Peet was entirely comfortable with this outcome: he had a small but profitable company that met his personal standards. Later in life, he acknowledged some downsides: he had missed out on some trends and opportunities, and suffered from the stress that went with total control.

Alfred Peet’s emphasis was on practical hands-on training. He felt cupping and roasting can be learned, but not from textbooks. Remarkably, he did not consider coffee to be the ultimate sensory challenge. For him, that was tea.
“In essence, tea is a more interesting product than coffee,” Alfred said. “Tea you can truly compare to wine, there are as many teas as there are wines, with as many different flavors, colors, and bouquets. (. . .) In contrast with tea, coffee is a rather coarse product; the subtle differences that you see in tea, you can’t find in coffee.”
Peet had little use for pompous, wine-like descriptions of coffee. He encouraged people to use simple, even personal terms to describe coffee flavor. His emphasis, especially when teaching someone how to roast beans, was on direct experience, on doing rather than knowing. This could lead to some unusual, almost Zen-like, directives.
How do you determine the moment when you have to take the beans out of the roaster? “The coffee will speak to you and will let you know when she’s ready,” Alfred told his students. For outsiders this sounds mystifying, but those who wanted to learn the profession from him developed an ear for it as well.
Peet didn’t invent a technology. He didn’t build a commercial empire or create a new way of doing business. Instead, he was a sensory entrepreneur: a blender of beans and an evangelist for flavor.

Alfred Peet has become known as the father of the American coffee renaissance. The impact of his life and work are evident wherever one goes. Today, there is quality coffee and espresso in crossroad burgs and small towns throughout the country’s vast interior.

For that, Mr. Peet, we thank you.

Thursday, August 16, 2018

Hard Copies & Web Metrics



Our cannabis aroma paper in PLoS ONE has now been viewed over 4,000 times and downloaded more than 600 times since February. It’s gratifying to see so many people interested in a new research theme. And yet, because web-based metrics are automated, impersonal counters, I don’t have a clue who these people are. It hasn’t always been this way.

When I was in graduate school in the pre-internet age, the super-efficient way to scan the new scientific literature was to grab a physical copy of Current Contents, the weekly publication of the Institute for Scientific Information (ISI). It was a thick little booklet, printed on thin, bright white paper. At least it was bright white when it arrived in the mail and when your lab head got to leaf through it. By the time it reached the grad students’ offices, it was tattered, marked up, and usually stained with coffee, cheesesteak drippings, and rat urine.



Current Contents listed the contents of the life sciences journals published that week, along with author names and addresses. This enabled you to request a physical reprint of the paper by mailing a letter or postcard. Better funded departments had pre-printed postcards for this purpose. Really well-funded labs signed up for ISI’s pre-printed Request-A-Print® cards—they came with a peel-off return address sticker for the reprint sender’s convenience. But you still had to fill in by hand the citation and the author’s address.



Although it beat going to the library and physically browsing recent, unbound issues of your favorite journals, this process was labor intensive. As a result, it made you think twice about the article in question: was it truly worth the effort of requesting your own personal copy? How many of us today download papers on a whim, only to have them stack up, unread, in our “worth a look” folder?

On the flip side, receiving reprint requests in the mail was a rewarding experience, especially for a graduate student or newly minted faculty member. It meant someone had found your work sufficiently interesting to fill out the postcard. And you knew who they were! A reprint request from a Big Name in the Field was acknowledgment that you had arrived.



Reprint requests also brought a sense of connectedness with scientists around the world. The mail would bring postcards with bright stamps and formal, cursive handwriting. The ones from communist block countries like East Germany, Poland, and Czechoslovakia were sad: printed on pulpy, low quality paper that oxidized in the sunlight and had a bitter smell. I always enjoyed fulfilling these requests—it meant acknowledging these scientists, keeping them connected to the free world, and encouraging freedom of inquiry.

Thursday, June 28, 2018

Cannabis in the U.S. Patent Office

Given my current interest in the olfactory properties of cannabis, I monitor the relevant scientific literature and every so often scan the U.S. patent application database. There are a ton of weed-related patents being filed. Some of the inventions are for practical devices, like this one for “Apparatus and related methods for trimming dried Cannabis flowers,” filed by Cullen Raichart of San Diego, California. Mr. Raichart aims to be the Eli Whitney of cannabis. The brief abstract and an illustration are all you need to understand his machine and what it does.


Then there’s “Apparatus and methods for biosynthetic production of cannabinoids,” by Robert Winnicki, et al. This group of inventors is looking beyond the agricultural horizon and proposing to create THC and related cannabinoids in a bioreactor. Neato!

As I browsed along, I found another application that sounded promising: “Method of producing cannabidiol derived products.” When I read the abstract I did a spit-take.
Methods are provided for producing cannabidiol (CBD)-based products in accordance with cycles of the moon and a plurality of biodynamic rituals. In some embodiments, the phase of the moon may be correlated with positions of the moon among the twelve zodiac signs, and the positions used to guide in the cultivation and harvest of cannabis plants. In some embodiments, the phase of the moon and the corresponding astrological phase that the moon is moving through may be used to determine when to plant seeds, when/how to water and fertilize, when to take cuttings, and the right time to harvest mature cannabis plants to produce optimal CBD-derived products. It is contemplated that farming cannabis with consideration of moon cycles may result in healthier plants, increased yields and superior products.
Yes, you read that correctly! Inventor Christine Meeusen of Merced, California, seeks to patent a method of using lunar cycle phase and the moon’s zodiacal position to optimize timing of seed planting, watering, and harvesting of Cannabis.

If the abstract sounds nutty, the body of Ms. Meeusen’s application (which cites no other patents or prior art) is even nuttier. Here’s item 37 from the Detailed Description section:
In one embodiment, the CBD derived products are only manufactured from the new moon to the full moon. On the first day of the cycle, a biodynamic ritual may be held under the stars so as to bless a work environment, such as a table. The ritual may include appreciation with respect to a Creator God and Mother Goddess. Depending on the product, a bundle of sage may be lit, and swept over any plurality of bottles and ingredients that include a high-proof alcohol to dissolve the bud. A plurality of incantations may also be communicated. In one embodiment, the producers may utilize a bundle of sage so as to cleanse one another prior to production. In one embodiment, appropriate dosages of marijuana may be scooped into preparation jars.
Ooooookay, then. Meanwhile, the flowchart in Fig. 1 looks like a game plan drawn up by the Underpants Gnomes:


Fig. 3 achieves peak nuttiness:


So who is Ms. Meeusen? A quick google search reveals that she is also known as Sister Kate the “weed nun.” Like other members of her “order,” she dresses in full habit, coif and wimple while running a sizable marijuana grow in Merced. According to Reuters, she first took up the nun persona in 2011 as part of her involvement in Occupy Wall Street. [This is my shocked face.—Ed.]

Is this a practical joke? Unlikely, given the expense of filing a patent. Is it performance art? Maybe. In any case, pity the poor patent examiner.

The U.S. patent applications discussed here are 20180126578 (Raichart), 20180179564 (Winnicki et al.) and 20180169162 (Meeusen).