Demetri Pananos

Demetri Pananos

Senior Data Scientist at Zapier. Part time PhD Candidate in Biostatsitics.

Blog posts

2022

Flippin’ Fun!

13 minute read

Published:

I wrote an answer to a question about sequences of coin flips a couple days back that I was quite chuffed with. In short, the question asked for statistical ways to determine if a sequence of coin flips was from an unbiased coin or a human trying to appear random. The resulting model turned into a fun game on twitter centered around determining if people who follow me could simulate a sequence of coin flips that looked random (without using a real RNG, or some funny workaround. I have a lot of faith in my twitter followers…maybe too much).

An Interviewing Retrospective

11 minute read

Published:

I’ve just recently accepted a new role as a data scientist at Zapier starting May 2022. This concludes a sequence of interviews on, what felt like, every other day. There’s a lot I learned in this experience, and I’d like to share some of that with you. I’m going to omit names of companies where needed, but I’ve tweeted enough for y’all to know who was in the mix for this role. At the end, I hope you learn something about yourself, your career, and interviewing.

An Interviewing Retrospective

11 minute read

Published:

I’ve just recently accepted a new role as a data scientist at Zapier starting May 2022. This concludes a sequence of interviews on, what felt like, every other day. There’s a lot I learned in this experience, and I’d like to share some of that with you. I’m going to omit names of companies where needed, but I’ve tweeted enough for y’all to know who was in the mix for this role. At the end, I hope you learn something about yourself, your career, and interviewing.

New Year, New Me (Who This?)

3 minute read

Published:

I’m not usually the one to make resolutions, but this year is a little different for a few reasons. I’m working full time with no intent to chicken out and go back to academia (despite my constant threats to no one in particular to open a café), which means I have money for the first time in a long time as well as some unstructured free time. Because I turn 30 this year, I thought it would be prudent to make some resolutions so that I do not coast through the next decade of my life (a complaint I continually made in my 20s).

2021

Gym Data

4 minute read

Published:

Prior to the pandemic, my school’s gym would tweet out how many people were in the weight room at a given time. This was useful for an instantaneous understanding of how busy the gym was, but was not useful if you wanted to answer how busy the gym might be later. During my PhD, I managed to create a little python script to scrape some tweets, extract the number of people in the weight room, and model the data with a linear model. The model did OK, but I wanted to take it a step further and estimate the effects of days of weeks, months, special events, and even weather (because, as the gym’s manager told me when I met with her, the weather is a major factor in people going to the gym).

In Response To Vaccine Hesitancy From Tommy Caldwell

9 minute read

Published:

On September 21st, 2021 I received an email from Tommy Caldwell, author and owner of Hybrid Fitness in my hometown of London Ontario, with the subject line “COVID-19, Kids, and Vaccines: A difficult decision for parents to navigate”. The email was sent from Hybrid Fitness’ email address (which I had subscribed to sometime in 2018, in the first year of my PhD). In that email, Mr. Caldwell acknowledges that vaccinating is a “no-brainer decision” and professes he is an inexpert in COVID19, science, and research, yet continues on to interpret statistics from a clinical trial intended to evaluate the vaccine against COVID19 in persons aged 12-15 years.

Three Months Into Industry

6 minute read

Published:

I “left” (I say left in quotations because I’m still working on my Ph.D part time) approximately 3 months ago. That is a bit of a milestone. It is a quarter of a year working at a senior-ish level as a data scientist at a national bank. I know a lot of PhDs, especially in quantitative disciplines, are thinking of making the jump from academia to industry. This sequence of blog posts is not advice on how to make that jump, but rather to document one perspective on what that change entails.

Answering Easier Questions

11 minute read

Published:

The 95% in 95% confidence interval refers not to the probability that any one interval contains the estimand, but rather to the long term relative frequency of the estimator containing the estimand in an infinite sequence of replicated experiments under ideal conditions.

2020

Don’t Select Features, Engineer Them

12 minute read

Published:

Students in the class I TA love to do feature selection prior to modelling. Examining pairwise correlation and dropping seemingly uncorrelated features is one way they do this, but they also love to fit a LASSO model to their data and refit a model with the selected variables, or they might do stepwise selection if they are feeling in the mood to code it up in python.

Intuitive Formulae Are Not Always Right

4 minute read

Published:

In the data science class I help TA, we’re going over confidence intervals. Thanks to the central limit theorem, we can report confidence intervals for the out of sample generalization error. Let’s assume our loss function is mean squared error. A confidence interval would then be

3 Rules For Giving a Sh!t

4 minute read

Published:

I spend a lot of time on cross validated (CV). CV is my statistical escape from statistics, and it is also a place where I like to prove to myself that I am good at what I do.

Log Link vs. Log(y)

2 minute read

Published:

You wanna see a little gotcha in statistics? Take the following data

What Do Monty Hall and A Steak Have In Common? I Marinate Both

5 minute read

Published:

It took me half a PhD to finally understand the Monty Hall Problem, but I think I get it now. I’ve been tutoring a very bright student in probability (which is so ironic given I failed it way back when) and I’ve surprised myself with how effective I’ve been at solving little homework problems I would have previously not been able to solve. It isn’t like my PhD has been full of these sorts of problems and by virtue to their exposure I’ve gotten better at them. I’ve just been thinking about stats longer than I was thinking about stats in undergrad which sort of leads to this unconcious thinking about homework problems. More on that in the end.

Nothing is Normal So Don’t Worry About The T Test

3 minute read

Published:

I hate the objection “I can’t use the t-test, my data aren’t normal”. I see it all the time on Cross Validated when a data analyst is tasked with analyzing an experiment after it has been performed. They have piles of data, thousands of observations, and they have no idea what to do with it. They know of the t-test, but they erroneously believe (through no fault of their own) that the t-test is only valid if their data are normal.

2019

GSoC 2019: It’s Over!

6 minute read

Published:

It’s the end of August, and Google Summer of Code 2019 is over. This blog post is meant to outline what I’ve accomplished, what I’ve failed to accomplish, what I’ve learned, and how I’ve felt over these last 4 months.

GSoC 2019: A PR is Made!

less than 1 minute read

Published:

Another short update: I’ve made a PR to merge pymc3.ode into PyMC3!

GSoC 2019: Testing an API

1 minute read

Published:

This is a really short update. Here are a couple things I have been working on since the last blog post.

GSoC 2019: Designing an API

4 minute read

Published:

Let’s take stock of where we are on our journey to add ODE capabilities on PyMC3.

GSoC 2019: Gradient Descent for ODEs (But This Time, In Theano)

8 minute read

Published:

A little while ago, I wrote a post on doing gradient descent for ODEs. In that post, I used autograd to do the automatic differentiation. While neat, it was really a way for me to get familiar with some math that I was to use for GSoC. After taking some time to learn more about theano, I’ve reimplemented the blog post, this time using theano to perform the automatic differentiation. If you’re read the previous post, then skip right to the code.

Gradient Descent for ODEs

7 minute read

Published:

Gradient descent usually isn’t used to fit Ordinary Differential Equations (ODEs) to data (at least, that isn’t how the Applied Mathematics departments to which I have been a part have done it). Nevertheless, that doesn’t mean that it can’t be done. For some of my recent GSoC work, I’ve been investigating how to compute gradients of solutions to ODEs without access to the solution’s analytical form. In this blog post, I describe how these gradients can be computed and how they can be used to fit ODEs to synchronous data with gradient descent.

GSoC 2019: ODE Inception

5 minute read

Published:

Let’s take stock of exactly where we are in this journey it implement HMC for differential equations.

2018

King Street Pilot Project

less than 1 minute read

Published:

Toronto started a pilot project to shut down King street to private vehicles in and attempt to ease congestion and increase TTC ridership. I’ve obtained some data from the city and have begun analyzing it. Shown here is an initial plotting of the change in travel times in certain sections of the city. Cooler colors mean travel times have decreased.

Neat Little Combinatorics Problem

2 minute read

Published:

I’ll cut right to it. Consider the set $S = (49, 8, 48, 15, 47, 4, 16, 23, 43, 44, 42, 45, 46 )$. What is the expected value for the minimum of 6 samples from this set?

Rat Tumors and PyMC3

less than 1 minute read

Published:

I’m very proud to say I have contributed this example to PyMC3’s documentation. It details how to compute posterior means for Gelman’s rat tumour example in BDA3.

Making Plots with Purrr

4 minute read

Published:

I was recently asked to make 4 plots for a collaborator. The plots are all the same, just a scatter plot and a non-linear trend line. Every time I have to do something repetitive, I wince, especially with respect to plots. I thought I would take this opportunity to write a short blog post on how to use functional programming in R to make the same plot for similar yet different data.

I C What You Did There, Sklearn

4 minute read

Published:

Let me ask you a question: Considering logistic regression can be performed without the use of a penalty parameter, why does sklearn include a penalty in their implementation of logistic regression? I think most people would reply with something about overfitting, which I suppose is a reasonable answer, but isn’t very satisfactory, especially since the documentation for sklearn.linear_model.LogisticRegression() is awash with optimization terminology and never mentions overfitting.

Advent of Code: Question 2

1 minute read

Published:

Today was a fairly easy challenge. Part one provides us with a 2d array of integers and asks to find the sum of the differences between the largest and smallest numbers in each row. Super easy to do without loops if you know how to use numpy.

A Purrrfect Method for Simulating Data

4 minute read

Published:

When I was doing my Masters, I had to generate a lot of plots, which means I had to generate a lot of data. Usually, the data I would be generating would depend on a parameter (maybe something like the rolling window length, or maybe the bandwidth for some smoothing function) and I would have to try a whirlwind of combinations. In order to do this, I would end up doing is writing code to generate the data once, then just loop over that code for different values of the parameters.

GyMBo: A Gym Monitoring Bot

less than 1 minute read

Published:

Back in September 2017, I was really tired and learning about Maximum Likelihood and yearned to do some more machine learning. For the longest time I wanted to scrape my gym’s twitter account to get information about when the gym was busiest.

2017

Coins and Factors

3 minute read

Published:

I love Fivethirtyeight’s Riddler column. Usually, I can solve the problem with computation, but on some rare occasions I can do some interesting math to get the solution without having to code. Here is the first puzzle I ever solved. It is a simple puzzle, yet it has an elegant computational and analytic solution. Let’s take a look.

The Very Start

less than 1 minute read

Published:

This is my humble website where I post math/stats/data science related stuff. Keep on the look out as I keep updating it with cool projects, questions, and thoughts.