Welcome to #DataIsForDoing

March 2023 Issue

Following the examples of Beyoncé, The Beatles, and Bananarama, all of whose eponymous albums are NOT their débuts, I decided that the third edition of the newsletter would be the one entitled "#DataIsForDoing." And we'll see if this newsletter can follow the success – either spectacular or dubious – of the artists providing my inspiration.

With that in mind, I'm so glad that you joined me here. I have a few goals for this newsletter that I'd like to share with you as a way of putting everything into its proper context. But, first, I should explain the title, , and so I'll share a (very) short video (and I apologize for YouTube's refusal to show the custom thumbnails on videos like this one that qualify as "YouTube Shorts." And if you're on a mobile device, you can see the video at this link: https://youtu.be/6qKm-gYkevo)

My goals for this newsletter

My hope is to help people use data to solve practical problems, as well as solve the practical problems that data work itself presents. There are a few approaches that I want support in the editions of this newsletter:

• Support an action-oriented approach to data work, where the goal of the data analysis is to guide people's behavior, as opposed to, say, theory building or factoid generation, as fun as those things may be
• Support efficient data work by choosing the simplest possible approaches to create "minimal viable analyses," or the statistical analyses that provide the needed insight while making the fewest demands on analysts or stakeholders
• Support human-centered data work: data should help people, not the other way around, and so data work should keep human goals and processes front and center, especially when it comes to interpretability and applicability

In fact, I should say a little more about that last point. I take a profoundly humanistic approach to data work. So, while data-driven (or data-informed) processes are important, there is both room for and a need for human experience and human judgment. Your projects need to make sense. Your conclusions also need to make sense. You should use your own experience and personal insights as a foundation for checking your analyses. I like to think of this as an intuitive form of Bayesian analysis, where the prior probabilities from your own experience are combined with the current analyses to yield more realistic and generalizable posterior probabilities. (And that's an obvious topic for a future newsletter, so I've already added it to my list.)

I'll mention one very important caveat about using one's own experience in data analysis. I would hope that it's obvious by now, but different people have different experiences, and no one experience is "authoritative." I've worked with people who were profoundly different from me – different continents, different languages, different cultures, different education and professional experience, different goals – and it would be foolish for me to use my experience for interpreting data from their context. And that, of course, is one reason why collaborative data work is so important if you want useful, relevant results. (I have also learned that my experience as an academic researcher doesn't always transfer well from my academic field to others, and that non-academic work is another thing entirely. I believe that I have adapted well to data work outside the Ivory Tower, but it took some time. That, too, is a topic for another day.)

A human audience

I need to make a distinction that may seem a little peculiar: my intended audience is humans, not machines. That is, I am concerned about data work that is to be conducted and used by humans to make decisions about the things that are important to them. It may not even be immediately apparent that there is an alternative to this focus, but I'm drawing a line between human decision-making and automated machine-learning processes.

I'm not criticizing machine learning. ML is an amazingly capable approach: dictation is a great tool (when it works), cars that can (sort of) drive themselves, ML-assisted medical diagnoses, the weirdness of DALL-E, and the shocking abilities of ChatGPT and other large language models. They're amazing. But, on a head-count basis, these make up only a small number of the data projects out there. For each splashy AI project, there are thousands (or even millions) of people, businesses, and organizations trying to decide, perhaps, if they need to stay open later on Thursdays, or how many postcards they should send out, or whether they should hire another employee. These are decisions made by humans and that are implemented by humans. It's a slower, smaller, and less glamorous process, but it's also what makes (the great majority of) the business world go round.

Some of the courses that I make for LinkedIn Learning focus on machine learning and artificial intelligence, but my main interest – especially here – is data work at human scale. In my consulting work, I have seen how important it is for people to get good answers to guiding questions:

• Goals. What, exactly, is the purpose of the project they are working on? How will they know that it will provide a useful answer to a meaningful question?
• Data gathering. When should they gather new data? How should they gather it? How much do they need?
• Analytical methods. What is the most direct way to get actionable insights from their data? When is a simple approach the best approach? And when are more complicated methods necessary for actionable insights?
• Presenting effectively. How can they best present their results and recommendations? How can they make sure the message got across properly?
• Efficiency. How can they save time and effort on future projects and work well with collaborators?

And so these are the kinds of topics that I hope to cover in the newsletter. Really, I want to help you use your time and your talents well, avoid some potential problems in your projects, and get meaningful, actionable insights with minimal stress.

Who I am

Finally, I wanted to take a minute to introduce myself. First and foremost, I am a teacher. I started teaching while I was still in graduate school, and I have taught full-time ever since I finished school. My formal training is in social psychology, which is a research field and not a clinical field. I have a bachelor's degree in psychology from Brigham Young University, a master of arts degree in psychology from Hunter College, and both a master of philosophy degree and a doctor of philosophy degree in social psychology from the Graduate Center at the City University of New York.?

But my process wasn't linear. I originally studied design. From as early as I can remember, I wanted to be a designer. I drew all the time as a child, and I took classes like color theory and 2D design in college. That was in the 1980s, before I had a computer, so I worked at a drafting table, drew with blue pencil on vellum, and made models out of foam core by hand. It was wonderful. I also did an informal internship at General Motors and found that one of the great things that can come from internships is knowing what you DON'T want to do. That experience convinced me that I didn't want to be a designer in a huge corporation. As a result, I switched to psychology. Design remains important to me, and I have been able, at times, to continue my creative work (including some wonderful collaborations with my wife, Jacque, who is a modern dance choreographer), but now my creative energy is mostly channeled into my teaching of statistics and data science.

Strangely, while I have four degrees, none of them is in statistics or data science. In those fields, I am essentially self-taught. A person I once met talked about the value of "autodidact swagger," or the confidence and resourcefulness that come from being able to learn things on your own. I'm a true believe in that approach, which is one of the reasons that I enjoy working with LinkedIn Learning. But it's not an exclusionary focus: I have been a full-time professor at Utah Valley University for twenty years, and I have also taught at Brigham Young University, the University of Utah, Salt Lake Community College, Lehman College, and Hunter College. I'm a believer in formal, higher-education, as well. I love both approaches.?

I'll finish with a few other random bits, just in case it helps put some of this work into context. I grew up in Los Angeles, went to college in Utah, spent two years in France, went to graduate school in New York City, and have been in Utah for the last 25 years, mostly in Salt Lake City. I love living in the West. I love riding my bike in Utah's West Desert, I love driving down the state to visit Zion National Park, and I will soon start rowing in the Great Salt Lake. As much as I loved living in Los Angeles, New York City, and Paris, I think the open space of the American West, as well as the practical nature of my ancestors who immigrated here, influences the way I think about things and solve problems, and I think that, in turn, influences how I approach data work, too. Geography may not be destiny, but its effects can be felt.

So, that's a short overview of my goals for , as well as some guiding principles and biographical details. More than anything, my goal here is to help you work productively with data so that you can solve actual problems without getting overwhelmed. I look forward to working with you. Thanks for joining me.

Thanks for joining me here. And remember, sharing is caring!?Follow me on LinkedIn?and share this??newsletter with a friend who you think would benefit from it.