Baker Moran's Blog

About this blog

Mon, 19 Jul 2021 15:41:47 GMT

I've long avoided creating a personal website, let alone a blog. I think I always viewed it as unnecessary for myself, and that I didn't have anything important to say. But I've since realized that I enjoy talking about things I find interesting, and it's ok if I write it down no one even reads it. The act of writing helps to evolve my ideas and helps me learn.

So why blog?

I've decided that its ok that the purpose of this blog is going to be mostly self-indulgent; I don't get to do much writing anymore, so this is a way to do so. I'm not going to promise that by reading this your worldview will be altered for the better, or you'll learn some shiny new work-related things. Basically, I'd just like to have a place where I can write some thoughts down and potentially have a discussion with anyone who happens to see it.

Writing can be a joyless activity when it's imposed by the powers that be (I hated writing papers in high school). In college, majoring in engineering, I was required to do very little writing, and the writing I did do (mostly lab reports, etc.) was incredibly boring to me. Post college, I've found that when I choose to write, and I'm allowed to do it in my voice rather than the robot voice required by the college of engineering, I actually enjoy it. While I don't expect anyone else to read this, if anyone in the black hole to which I am sending these posts called the internet ends up reading and enjoying them, I'll consider that an unexpected yet welcome outcome. I've found that the best way for me to evaluate and evolve my own thinking is to write it down, so if nothing else, I hope only to accomplish this.

A note on style

Because this is mainly for myself, I'm going to write how I like to write, which is how I talk. In my experience, people find it easier to read things that are written in the way they're most familiar with. So if you don't like run-on sentences, comma splicing, and informal language then this is not the place for you. If you don't like attempts at humor while discussing serious topics, then this is not the place for you. If you like to take a lighthearted approach to learning something new and questioning the way that you view the world, then you've come to the right place.

Topics I'll write about

I sort of pride myself on having an extremely broad yet probably shallow range of knowledge. I have the most fun when I'm learning about a topic that I know little about, but when I reach a minimum level of competency on that topic, I usually abandon learning anything else about it, but I rarely forget what I initially learned. Because of my love of extremely varied topics, expect these blog posts to touch on a wide range of subjects going into a small amount of depth in each post. My goal with this approach is threefold: 1. alienate any possibility of a regularly interested base of readers by ensuring that they can never come back to read more on a specific topic, 2. anger everyone that knows more about the topic than I do; since this is the internet, that is guaranteed to happen and 3. more seriously, potentially introduce others to an idea or topic that they've never encountered, and at best change the way they think about the world.

See you soon

So with that said, I will hopefully set aside time to write in the near future. I look forward to writing about things I care about and getting to talk to new people! If you want to connect, please don't hesitate to reach out!

Cities are more expensive than ever

Fri, 01 Jul 2022 01:45:28 GMT

In the midst of covid, many predicted the “death of in person work”. To a large extent, this has been true - many companies have chosen not to go back to the office and have instead started hiring workers from all over the country. Why then do we see rents in places like New York City and San Francisco skyrocketing to higher than pre-pandemic levels?

While it's true that rents initially dropped at the beginning of covid as people moved out of cities amidst the uncertainty, they have stormed back in numbers even greater than before as they come to terms with what moving away was like. When everyone is able to work from home, why would someone choose to live in a 500 square foot, $3,000 a month apartment in Brooklyn when they could live in a house more than double the size and less than half the cost in a midwestern suburb?

I think the answer to this question lies in something deeply true about human nature. People need to be around other people. They desire to have easy, walkable access to things like restaurants, bars, and grocery stores. These things demonstrably improve quality and happiness in our lives. Now more than ever it’s clear that the car centric suburban experiment we’ve been sold as the American Dream for the last 75 years (which was legislated into reality) is an abject failure.

But the tides are turning, people are realizing that they are better off living in denser communities, close to their daily needs. They are moving back to high population cities, and rents are skyrocketing. They are moving from the suburbs and buying up real estate in the denser urban cores of midwestern cities. This trend is nothing new, but like so many things, it's been exaggerated by covid - after a year plus, people have realized the void that isolation left in our collective lives, and how much better it can be . We have always innately known[^1] that living in close-knit and walkable communities is more life-giving, even when we’ve been told the dream is to live in the suburbs.

More often than not (especially with current gas prices), when you’re choosing between staying in or driving 20+ minutes to a bar somewhere to deal with parking, you’ll choose to stay in. When the options are instead to walk out your door to a diverse range of local restaurants or to choose to stay in, you’ll choose to go out and be around people. It really is that simple. People are willing to pay an extreme premium, both monetarily and in quality of housing, to be close to interesting things. They are not just willing to put up with some of the downsides that come with density in most American cities in order to live in a walkable area, but choosing to more than ever, even in the remote work era.

But therein lies the problem. Currently, due to years of poor zoning practices, living in this type of mid-density community is outrageously expensive. It’s out of reach for most people, so they have to live in car-dependent suburbia; there is no other option. Shouldn’t this cause us to think? Shouldn’t this cause our city planners to revisit our insane and outdated building codes that prevent these developments and lead us to reform our zoning across the country? The mixed-use medium and high density housing most people long for is effectively banned in all but a few places[^2], leaving it out of reach and prohibitively expensive for all but the wealthiest Americans. But clearly, the demand is there. People want this type of development. It doesn’t have to be this way. We shouldn’t have to choose between just a few cities that offer livable communities. We shouldn't have to pay most of our paychecks to live in a walkable area. We deserve better.

[^1]: There is a reason that most people who went to college view it as one of the best times of their life - it is the only time in many people’s lives when they experience the kind of proximity to friends, work, restaurants, and everything else they need that most college campuses offer.

[^2]: It only exists in a handful of cities that didn’t have the chance to bulldoze their dense urban cores for parking lots when cars become commoditized.

Why use Bayesian Statistics for AB Testing

Wed, 21 Jul 2021 19:48:04 GMT

There are plenty of articles out there motivating why we should use Bayesian statistics in product analytics. I'm not going to try to reinvent the wheel here, so I will provide only a brief explaination and link to some articles that I found helpful in my learning. I will also explain why I created this package.

Motivation

Why Bayesian Statistics

Bayesian statistics is a completely different way to think about statistics from what you likely learned in high school and college (typically referred to as frequentist statistics). Bayesian statistics allows us to much more easily answer a wider variety of questions that are more relevant for the business world. Although most people are used to hearing a p-value for a statistical test, many don't actually know what it means or how to interpret it beyond the "satistical significance is when p <= 0.05" we were programmed to regurgitate. Bayesian statistics allows a much more intuitive interpretation of the results of a test. Examples of questions Bayesian statistics is purpose built to answer:

What is the probability that variant we are testing is better than the control?
How much better is the variant than the control?
If we make the wrong choice, how much worse off might we be?

The Issue with Traditional AB Testing

The goal of using frequentist statistics is to minimize the probability of being wrong when we pick the variant over the control. P-values are designed to be biased towards the control. In business we often run an experiment because we believe we are making an improvement to the product. There clearly needs to be statistical rigor, but a question I often get when the variant is slightly better than the control, is "why can't we just pick the variant?". Bayesian statistics allows for statistical support, even when picking a variant that is only slightly better.

Frequentist statistics protects us against choosing something new that isn't actually better. This is important in things like medicine; it's not that important in the business world. In business, we want to run lots of tests as quickly as we can, in order to make the best decisions we can about the business. Changing the color of a button on the website likely will not result in lives lost, while a new medication could. Bayesian statistics allows us to control the risk we are taking on every decision we make; we can choose to make a decision with less data than we would need with Frequentist, while controlling our risk.

The Power of Bayes Theorem

Bayesian statistics is designed to use our belief about the world in order to help us make a decision. At first, this was confusing (at least to me) because it sounded like an arbitrary choice. And it is; but the key to understanding why Bayesian statistics is so powerful is that when using frequentist statistics, we are making even stronger and more arbitrary assumptions. In business and tech, we often have access to a lot of data and have a pretty good idea about conversion rates. I would argue that not using any of that information is a more egregious mistake than using the wrong prior with Bayesian statistics.

Goal of the BayesABTest Package

I created this package originally out of necessity and a desire to learn.

I previously had little exposure to Bayesian statistics, and I found myself wanting to learn about it. The best way for me to learn things is to teach other people and build something.
The code we used at Root for testing only included functionality for one variant conversion data AB tests. I often needed the ability to use different priors for continuous data, as well as additional variants.
I realized there are several very robust R packages for Bayesian AB testing (I'll link a few), but I would much rather use Python than R, and I couldn't find any good Python alternatives. For this reason, I decided to expand and productionalize this package into something that analysts at other data driven and testing oriented companies can use.
This package is meant to be an easy to use method for analysts to report AB test results in a Bayesian framework using Python.

R packages that inspired BayesABTest

bayesAB - bayesAB by Frank Portman
abayes - abayes by Convoy Inc.

Why you need to learn another language

Wed, 20 Oct 2021 01:30:29 GMT

Learning a new language is something that I believe can offer you a unique and different perspective on the world and alter the way you think. It provides a framework to critically think about the way you comprehend and perceive the world around you. Learning a new language has altered my view of the world for the better, and it can do the same for you.

I’ve discovered over quarantine that one of my passions is language. I’ve always had a passive interest in linguistics but never delved into it much. This past year I’ve picked up learning French while stuck at home; initially, it was mostly about feeling like I had something productive to do while I sat on my phone for hours while stuck at home. But it has since morphed into a genuine passion.

Why should you care?

I’m sure you’ve heard this many times before, both in school and in life: “learn another language”. We as Americans are slacking when it comes to speaking a second language[^1]. The question for most people is, why does it matter? Some responses you’ve likely said or thought:

English is the lingua franca (common language) for most of the world, and I already speak it
I am never in a situation where I need to speak another language with someone
If people come to America, they should learn to speak English

I could refute each one of those points; “It's a sign of respect to others”, “it could allow you to travel more easily”, “the US has no official language”, but at the end of the day, arguing with someone rationally on these points will never motivate them to learn a new language. It’s important to understand that learning another language is a large undertaking, and as many adults have learned if you’re not forced to learn something like piano or Spanish as a kid, it’s not going to happen as an adult.

How to cross the mental hurdle that “it’s hard”

Something I’ve learned and really taken to heart in my short professional career is that the key to getting people to do something that you want is all about incentives (hoping to write about this soon). People have to think that they will gain something by doing the thing you’re asking of them. To get a large number of people to do what you want, just give them a selfish reason to do so.

Learning a language is a large amount of work, and for an American that is a native English speaker, I can’t refute that there’s almost no immediate payoff. The things I mentioned above have no intrinsic value to someone learning a language; so, I will attempt to convince you that there is knowledge out there waiting to be gained.

Expanding your world

The way you grew up and learned things affects the way that you perceive the world. You react and perceive things a certain way, without ever intending to. For example, you call the sky blue because we have a word for blue, while we could just as easily have a color for a light blue as we do for pink (light red). The language you grew up speaking contributes to the way you perceive and interact with the world around you.

Implicit bias affects the way you see the world
Implicit bias is caused by your learned experiences
Your native language contributes to your learned experiences
Therefore, the language you speak affects the way you see the world

I think the best way to explain this fact is by way of a few examples. In the end, you may say great, none of this is true of English. However I ask this, how can we understand these biases and differences without understanding the alternatives? If no one ever learned another language, we would never know that anyone else sees the world differently.

We can see from these examples how growing up a speaker of another language could affect someone, but what do you, having never learned another language, not know about how being an English speaker affects the way you think? The simple act of learning a new language unlocks this. There is much knowledge to be gained simply by comparison.

Gendered Nouns

There have been multiple studies on the way that gendered nouns affect the way you perceive the world. In German, “bridge” is a feminine noun; in Spanish, it is masculine. In a 2003 study[^2], native German and Spanish speakers were asked to categorize a list of nouns. The catch was, all participants were fluent in English, and the study was conducted in English.

“German speakers described bridges as beautiful, elegant, fragile, pretty, and slender, while Spanish speakers said they were big, dangerous, strong, sturdy, and towering”[^3] all while speaking English. The fact that these participants described the words as such while speaking an un-gendered language shows the way that the genders subtly affect the way we think about things around us. Even the way you perceive inanimate objects.

Past, Present, and Future tense

Unlike English, Indonesian does not have verb endings that indicate tense (this is a hard concept to grasp for an English speaker). In a 2002 study[^4], participants who spoke English and Indonesian were shown pictures of a man who had either just kicked a ball, was in the process of kicking a ball, or who was about to start kicking a ball. They were then asked to later recall what was happening in the photo. The English speakers were able to recall which of the three they were shown with much greater accuracy. The language you speak and think in subconsciously influences even the way you remember your experiences.

Number of words

What if I told you that the way you view color is highly based on the language you speak. As English speakers, we have 11 main colors. In order to describe another color, we typically list one of those colors and modify it with light, dark, or something else. When you see something and are asked to describe the color, you will almost always describe it as one of those basic colors, even if it doesn’t exactly match.

There are languages around the world with different numbers of basic colors. Some languages have just two colors (light and dark). Would they call something that is dark-blue light-black? As an English speaker, you would say that’s just wrong; but to them, it would be correct! If you were to tell a Russian or Greek speaker what color English speakers say the sky is, they may have the same reaction! We call it blue, or light blue to be more specific. But Russian and Greek have their own basic color terms for light and dark blue, in the same way, we have red and pink while some cultures would just call pink light red. In some languages, green is a shade of yellow, or blue and green shades of the same color.

It’s often said that the more words a language or culture has for something, the more deeply they understand it. You’ve likely heard the common saying that many Eskimo-Aleut languages have 50+ words for snow, while we just have one. They may have words for sleet, powdery snow, wet snow, etc. We just have to describe it, but they intuitively know it by the language. When you learn another language, the differences between it and English may not be this extreme, but you have a new framework to think about the world when you start thinking from another perspective.

Takeaways

If you’re anything like me, these conclusions are mind-boggling. Something as simple as language has the ability to influence how we perceive the world, and how we remember our experiences. The question is, what do you not know about how your experience is being shaped because you are an English speaker? You have the power to unlearn what you’ve learned since birth, and unlock new perspectives, simply by spending a few minutes a day learning a new language.

[^1]: Over half of the world can speak two or more languages, and only 20% of Americans can. And let's be honest, most of that 20% likely comes from immigrants to America

[^2]: Boroditsky, L., Schmidt, L., & Phillips, W. (2003). Sex, Syntax, and Semantics. In Language in mind: Advances in the study of language and cognition, ed. D. Gentner & S. Goldin-Meadow, pp. 61- 80. Cambridge University Press.

[^3]: "Masculine or Feminine, and Why It Matters", Psychology Today.

[^4]: Boroditsky, L., Ham, W., & Ramscar, M. (2002). What is universal in event perception? Comparing English & Indonesian speakers. Proceedings of the Annual Meeting of the Cognitive Science Society, 24. Retrieved from https://escholarship.org/uc/item/5jd5w5c8