#274: Real Talk About Synthetic Data with Winston Li

Synthetic data: it's a fascinating topic that sounds like science fiction but is rapidly becoming a practical tool in the data landscape. From machine learning applications to safeguarding privacy, synthetic data offers a compelling alternative to real-world datasets that might be incomplete or unwieldy. With the help of Winston Li, founder of Arima, a startup specializing in synthetic data and marketing mix modelling, we explore how this artificial data is generated, where its strengths truly lie, and the potential pitfalls to watch out for! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

Is it just us, or are data products becoming all the rage? Is Google Trends a data product that could help us answer that question? What actually IS a data product? And does it even matter that we have a good definition? If any of these questions seem like they have cut and dried answers, then this episode may just convince you that you haven't thought about them hard enough! After all, what is more on-brand for a group of analysts than being thrown a question that seems simple only to dig in to realize that it is more complicated than it appears at first blush? On this episode, Eric Sandosham returned as a guest inspired by a Medium post he wrote a while back so we could all dive into the topic and see what we could figure out! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

No matter how simple a metric's name makes it sound, the details are often downright devilish. What is a website visit? What is revenue? What is a customer? Go one level deeper with a metric like customer acquisition cost (CAC) or customer lifetime value (CLV or LTV, depending on how you acronym), and things can get messy in a hurry. In some cases, there are multiple "right" definitions, depending on how the metric is being used. In some cases, there are incentive structures to thumb the definitional scale one way or another. In some cases, a hastily made choice becomes a well-established, yet misguided, norm. In some cases, public companies simply throw their hands up and stop reporting a key metric! Dan McCarthy, Associate Professor of Marketing at the Robert H. Smith School of Business at the University of Maryland, spends a lot of time and thought culling through public filings and disclosures therein trying to make sense of metric definitions, so he was a great guest to have to dig into the topic! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

Data that tracks what users and customers do is behavioral data. But behavioral science is much more about why humans do things and what sorts of techniques can be employed to nudge them to do something specific. On this episode, behavioral scientist Dr. Lindsay Juarez from Irrational Labs joined us for a conversation on the topic. Nudge vs. sludge, getting uncomfortably specific about the behavior of interest, and even a prompting of our guest to recreate and explain a classic Seinfeld bit! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

We finally did it: devoted an entire episode to AI. And, of course, by devoting an episode entirely to AI, we mean we just had GPT-4o generate a script for the entire show, and we just each read our parts. It's pretty impressive how the result still sounds so natural and human and spontaneous. It picked up on Tim's tendency to get hot and bothered, on Moe's proclivity for dancing right up to the edge of oversharing specific work scenarios, on Michael's knack for bringing in personality tests, on Val's patience in getting the whole discussion to get back on track, and on Julie being a real (or artificial, as the case may be?) Gem. Even though it includes the word "proclivity," this show overview was entirely generated without the assistance of AI. And yet, it’s got a whopper of a hallucination: the episode wasn’t scripted at all! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

How is an outlier in the data like obscenity? A case could be made that they're both the sort of thing where we know it when we see it, but that can be awfully tricky to perfectly define and detect. Visualize many data sets, and some of the data points are obvious outliers, but just as many (or more) fall in a gray area—especially if they're sneaky inliers. z-score, MAD, modified z-score, interquartile range (IQR), time-series decomposition, smoothing, forecasting, and many other techniques are available to the analyst for detecting outliers. Depending on the data, though, the most appropriate method (or combination of methods) for identifying outliers can change! We sat down with Brett Kennedy, author of Outlier Detection in Python, to dig into the topic! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page. 

Do you cringe at the mere mention of the word, "insights"? What about its fancier cousin, "actionable insights"? We do, too. As a matter of fact, on this episode, we discovered that Moe has developed an uncontrollable reflex: any time she utters the word, her hands shoot up uncontrolled to form air quotes. Alas! Our podcast is an audio medium! What about those poor souls who got hired into an "Insights & Analytics" team within their company? Egad! Nonetheless, inspired by an email exchange with a listener, we took a run at the subject with Chris Kocek, CEO of Gallant Branding, who both wrote a book and hosts a podcast on the topic of insights! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

Why? Or… y? What is y? Why, it's mx + b! It's the formula for a line, which is just a hop, a skip, and an error term away from the formula for a linear regression! On the one hand, it couldn't be simpler. On the other hand, it's a broad and deep topic. You've got your parameters, your feature engineering, your regularization, the risks of flawed assumptions and multicollinearity and overfitting, the distinction between inference and prediction... and that's just a warm-up! What variables would you expect to be significant in a model aimed at predicting how engaging an episode will be? Presumably, guest quality would top your list! It topped ours, which is why we asked past guest Chelsea Parlett-Pelleriti from Recast to return for an exploration of the topic! Our model crushed it. For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

  In celebration of International Women’s Day, this episode of Analytics Power Hour features an all-female crew discussing the challenges and opportunities in AI projects. Moe Kiss, Julie Hoyer and Val Kroll, dive into this AI topic with guest expert, Kathleen Walch, who co-developed the CPMAI methodology and the seven patterns of AI (super helpful for your AI use cases!). Kathleen has helpful frameworks and colorful examples to illustrate the importance of setting expectations upfront with all stakeholders and clearly defining what problem you are trying to solve. Her stories are born from the painful experiences of AI projects being run like application development projects instead of the data projects that they are! Tune in to hear her advice for getting your organization to adopt a data-centric methodology for running your AI projects—you’ll be happier than a camera spotting wolves in the snow! 🐺❄️🎥 For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.

Every listener of this show is keenly aware that they are enabling the collection of various forms of hyper-specific data. Smartphones are movement and light biometric data collection machines. Many of us augment this data with a smartwatch, a smart ring, or both. A connected scale? Sure! Maybe even a continuous glucose monitor (CGM)! But… why? And what are the ramifications both for changing the ways we move through life for the better (Live healthier! Proactive wellness!) and for the worse (privacy risks and bad actors)? We had a wide-ranging discussion with Michael Tiffany, co-founder and CEO of Fulcra Dynamics, that took a run at these topics and more. Why, it's possible you'll get so excited by the content that one of your devices will record a temporary spike in your heart rate! For complete show notes, including links to items mentioned in this episode and a transcript of the show, visit the show page.