PyData x PyYYC: Building an Intuition for Probability & Highly Performant IO

This is a special meetup, as we've joined forces with our compatriots at PyYYC for a joint meetup! We will have two talks, the PyData Probability Prerequisites and the PyYYC Highly Performant IO. Check out more PyYYC goodness at https://www.meetup.com/py-yyc/ for those interested in improving their programming and development skills.

Please note the time change to accommodate our friends at PyYYC.

PyData Talk:

In this introductory session, we will develop an intuition for some of the fundamental concepts of probability. This session is aimed at those with little to no exposure to probability and statistics who are interested in learning more about data science and statistics. It is also a prerequisite for our Bayesian Methods series.

We will develop intuitions for

• Stochastic (random) vs deterministic variables
• Independent, identically distributed random variables
• Marginal, joint, and conditional probability
• Expectation
• Bayes theorem

Experienced members of PyData can consider skipping this session, but are welcome to join to share their insights and mentorship. The content of the session will be shared as a notebook for future reference.

PyYYC Talk:

Peter Hunnisett:
Highly Performant IO

-----------------------------------------------------------------------------------------

We've joined the YYC Data Community on slack! Rather than creating yet another slack workspace, we've aligned with some of our other friends in the community to try and create a more shared workspace to connect with the community. Join us here https://join.slack.com/t/yycdatacommunity/shared_invite/zt-9g38lqk7-No1saecPXHvpwO3fMfEgFQ

-----------------------------------------------------------------------------------------

This seminar is a part of a larger group of talks focused on Bayesian statistics, primarily using MCMC. This session is a prerequisite to ensure everyone is able to follow along as we progress.

---- About the Bayesian Methods series --------

The goal of the Bayesian Methods series is to empower the members of the PyData community to understand and apply the most common techniques of Bayesian inference and can begin to accommodate the techniques into their workflows. At the end of the series, users will be able to:

• Apply Bayesian inference for parameter estimation
• Use Markov Chain Monte Carlo in pymc3 to build hierarchical models, GLMs, etc for real world applications
• Understand dependent sampling
• Transfer knowledge of bayesian methods to frameworks such as Stan and tensorflow probability
• Demonstrate basic use of Variational Inference techniques

Inspiration for the series comes from:
https://github.com/CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers
https://www.amazon.ca/STUDENTS-GUIDE-BAYESIAN-STATISTICS/dp/1473916364
https://docs.pymc.io/

The series focuses on understanding through computation and intuition, making the series more accessible to those without a strong math background. We won't neglect the math, but if you and calculus aren't on speaking terms, it shouldn't hold you back.

The events will be half seminar format, half workshop format. We will spend 30-45 minutes going through the main concepts then the remaining 30-45 minutes working through the examples and exercises. This series is intended to be interactive.