Machine Learning Workshops (Livestreams)

Learn about how overfitting and multicollinearity are problematic and how to think about them via the bias-variance tradeoff framework. Learn the intuition behind different distance measures (L1, L2, etc) and how Regularization using Ridge, Lasso and ElasticNet help reign in extreme model coefficients. Finally, combine regularization techniques with the log normalization of skewed variables to generate a final predictive model.

In this livestream we dig into modeling best practices for Linear Regression including: cross validation, filling null values, one hot encoding, feature scaling and using Scikit-Learn pipelines.

We take a detour from Linear Regression to cover the Gradient Descent algorithm in the context of simple linear regression which helps us to visualize exactly how gradient descent finds the optimal model parameters.

Get introduced to Simple Linear Regression. We'll start with a baseline model, learn about three important error metrics: MAE, MSE, and RMSE. And then we'll fit a regression using Scikit-Learn and talk about how linear regression calculates the parameters that minimize the sum of the squared residuals.

Build an increased mathematical intuition around the critical statistics concepts of: variance, standard deviation, covariance and correlation –with covariance and correlation being our first foray into bivariate EDA.

Learn about appropriate univariate data visualizations for both categorical and continuous data including: Histograms, Density Plots, Box Plots, Bar Charts, Stacked Bar Charts, and Pie Charts.

We lay the ground work for our journey through applied math for data science beginning with Exploratory Data Analysis with a focus on univariate EDA and data visualizations.

An overview of different job titles under the umbrella of "data professional" and an exploration of how much math preparation is expected for different roles.

A deep dive on the Decision Tree algorithm including: decision tree visualizations, gini impurity, how decision thresholds are calculated, and calculating the best threshold with Python.

Provides a framework for approaching Machine Learning. Explores the categories of classification and regression algorithms. Basic data cleaning and beginning modeling using a baseline.

Covers: Categories of ML algorithms including Supervised and Unsupervised learning. Important vocab like: model, algorithm, tabular data, feature, target, generalizability. Fitting a Decision Tree model with Scikit-Learn.