MY Data Science: April 2020

A man standing in front of a wall of beer cans wearing a black "Lost Forty Brewing" polo.

How data science could help create even better brews

Grant Chandler,
Quality Control Manager,
Lost Forty Brewing

I WAS PLANNING to be a research scientist, and then beer got in the way. But it’s not like you’re thinking. I grew up in Houston, Texas, and went to college at Hendrix in Conway, where I majored in biochemistry. After graduation, I thought about going to graduate school to study some kind of molecular biology. In the meantime, I took a job as a lab technician at the Arkansas Children’s Hospital Research Institute. It’s the research arm of the hospital.

Becoming a research scientist had been my working hypothesis for a long time, and now I was getting to test that hypothesis. It wasn’t long before I realized that being a scientist wasn’t for me. The pace is pretty slow in research. There’s a lot of banging your head against the wall. It’s intentional failure, where you go down this track and then cross that one off and start going down another track that is nearly identical with a very precise variable edit. Overall, I found it rather tedious work, which I kept doing for two and a half years.

To combat the tediousness, I took up several hobbies—gardening was one, and kickboxing was another. Still another was home brewing. It was the first time in my life that I’d had time to burn. Having a stable living from working at the lab, I was able to approach these hobbies as I had my studies—it was all very structured and disciplined, but not rushed: just trying them out to see how I liked them. And in the end, brewing took over. Brewing is the best mix and balance of all my interests—history, community, science, art, nutrition, etc. There’s a lot of creativity involved in brewing beer, and up until then I had found my life rather lacking a sense of creativity. I can definitely express things that I like and don’t like through the artistic aspect of beer.

This is supposed to be about beer and data science, and I’ll get to that. You collect a lot of data when you work at creating something, but I wasn’t thinking about data as data at first. When I started brewing beer, out of my house, it was a total lark. I was essentially just throwing darts at a dartboard. I knew nothing about brewing beer, though I did know about yeast and fermentation. So I read up on brewing while I experimented.

Because of my interest in microbiology and fermentation, it was kind of a natural course to start questioning what the limits were of the institution that is beer and the fermentation that most brewers consistently use, which is very traditional. They’re the kinds of cultures that brewers have been using for centuries, very domesticated strains of the yeast organism. What I wanted to know was, What’s different about brewer’s yeast and other yeasts? Why can’t this yeast make beer? Why does this other yeast make better beer? So I started collecting non-traditional yeasts to see what flavors and performance and fermentation I could create. To put this in perspective, a very common wild mixed culture is a sourdough starter, which is the same sort of methodology I was using. But instead of flour, I used barley sugar.

To make good beer, your yeast has to have an affinity for unfermented beer, which we call wort. So I just put wort out and experimented to see what fermented it, and if it yielded any good results. Sometimes I would isolate the culture, and sometimes I would keep it mixed. Then I would make a batch of beer with it and see what it tasted like. Sometimes they were good, a lot of times they weren’t.

Then in 2014 I got involved with Flyway Brewing Company in North Little Rock. I didn’t work for them, but Matt Foster, the proprietor, was involved with the Arkansas Native Beer Project, in which the idea was to make beer only from ingredients native to Arkansas. Matt basically needed someone to forage wild yeast and I was someone who was seeking an opportunity to do just that. And it was during that time that I found a strain of Saccharomyces yeast in the Dunbar Gardens and used it to brew a hoppy ale that I called Dunbar Wild, which won a competition. Soon after that I joined Lost Forty, where that Dunbar yeast has continued to be a performer to this day.


TO ME, A good beer is a simple beer. By that, I mean water, malt, yeast, hops, in just the right balance. Craft beer as it exists today is so subject to trends and whims and wow marketing objectives, and those beers tend to be quite inferior in my opinion. They put in all kinds of adjunct ingredients instead of focusing on process and fermentation and consistency. My first favorite beer was Guinness, and I very much like Pilsners, Saisons, Blonde ales, Dry Stouts. These styles exemplify the maxim, “Simplicity is the ultimate sophistication.” Mexican lagers are great—we make one here called Easy Tiger. We’ve actually won an award for it, and it’s a very simple beer.

I think beers that hail from the Belgian tradition really exemplify the craft of high-end simplicity. Belgian beers historically reflect very simple recipes and processes but are nevertheless bursting with character, and that kind of balance inspires me. In contrast, German beers are very engineered, very precise, very consistent; they’re spot on. But they can be kind of hollow and boring sometimes. They can still taste great, especially for certain occasions, but Bavarian beers tend to be full of Brains instead of Heart.

My title here at Lost Forty is head of quality control, but I do a lot of things. I oversee all brewing operations and make sure everyone’s dotting their i’s and crossing their t’s. I also oversee our quality lab and our propagation lab, where we grow all of our fermentation cultures. We run a lot of diagnostics and collect a ton of data here, because we measure so many things—pH, yeast counts, temperature, gravity, microbiological stability; concentrations of chemicals that are flavor active, off flavors in particular. We taste what we’ve made and look back at the data and try to hypothesize why it might contain too much of this or too little of that.

Much of my job involves teaching and training. The creative part is hard to teach, but not impossible if someone really cares about beer. If they don’t care about beer, they can maybe follow instructions but they’re not going to be invested in trying to make things better. That being said, I do my best to convey the “feel” of the brewing process. Over and above all of our data and our instructions and our rules and standard operating procedures, I work hard to convey a sense of in-tuneness with the flow, especially regarding the fermentation. As engineered as beer and breweries can be, when you’re dealing with a living organism there’s always going to be unpredictability, and that’s something I love. It’s not just a matter of “use this yeast/don’t use that yeast.” You’ve got to work with the yeast, flow with it. That’s something that I think is important for any good brewer to know how to do.

One thing I really appreciated about my home brewing hobby was the total freedom to fail. Here at work, we have a little bit of that privilege, but much less so. Lost Forty is a real business, not a hobby, which was part of what attracted me to it. And in the years since I’ve been here, I’ve developed a new interest: business. I’m presently working on my MBA at UA Little Rock, and part of what I’m hoping to get out of school is to be able to manage our growing amount of data better. We keep all these data logs on the computer, and there’s definitely a lot of “feel” that goes into the manipulation of our data. What we do isn’t trial and error—we refer to this data all the time regarding our intentions and results. But we’re not systematic at all. This may just be another way of saying that we lack scientific data analysis skills. I would like us to make that data work for us more.

For a long time, we haven’t had the data to do that, even if we wanted to. We’re only five years old now, and I believe we’re just getting to the point where we have enough data to generate some predictive information. But because we make so many different brands and intentionally do things a lot of different ways, we are always experimenting if you will, and it could be that for any given variable set there might not yet be enough significant data points to analyze.

Friends have asked me if our becoming heavily “data centric” would take the fun out of brewing beer for me, and that’s certainly something to think about. But I’m not convinced that science and creativity can’t live in harmony together. My sense is that we can be both super analytical while remaining loyal to the creative intuition and experimentation that has guided our craft so far. A whole yin-yang instead of just half.

That’s my new working hypothesis. Hopefully we can increase our data analysis skills through hiring or education soon and put it to the test.