Have you ever wondered if you are looking at a fluke result or part of a consistent pattern? Have you ever faced the challenge of making a decision with only limited historical data and felt uncertain just how representative it may be? Decision making under uncertainty is a fact of life in our complex and ever-changing environment. Quantifying uncertainty and using data to model it lies at the heart of statistical computing. This course is designed to teach you the practical skills you need to master statistical thinking and apply it to real business problems using Python.
In this course, attendees learn fundamental and advanced statistical techniques for hypothesis testing and analyzing variance in data. They gain a broad understanding of many types of probability distributions and when they should be applied. Participants will learn to approximate distributions using Monte Carlo methods and will learn Bayesian thinking, plus the skills to apply it using Markov Chains. Additionally, they will also get hands-on experience with core Python statistical computing libraries such as Statsmodels, Scikit-Learn, Numpy, Scipy, and PyMC.
What You Will Learn
This course will teach you to think about problems like a statistician, applying both frequentist and Bayesian approaches to decision making under uncertainty. You will learn about probability distributions, sampling and resampling, Monte Carlo methods, and Markov Chains. Additionally, you will learn how to estimate uncertainty around regression analysis using bootstrapping techniques. You will learn how to select and apply key Python libraries to tackle complex data problems.
This course will cover the following topics:
Calculating descriptive statistics
Frequentist statistical modeling and simulation
Monte Carlo simulation
Hypothesis testing using estimated and derived probability distributions
Developing customized bootstrap resampling techniques
Resampling and confidence intervals
Regression analysis and interpretation with Statsmodels
Bayesian statistics including calculating posteriors
Implementation of random walks and Markov Chains
Markov Chain Monte Carlos for simulating complex posteriors with PyMC
Advanced Bayesian statistics with Gibbs Sampling and Metropolis-Hastings
We will use the Anaconda (Python 3.6 version) distribution for this course. Attendees will need a computer with Anaconda successfully installed.