Supervised Learning - Regression Models

Explore diagnostic analytics by answering why events occur, using covid case trends to illustrate how factors like lockdowns and vaccination explain rises and drops.

Learn how predictive analytics uses current data to forecast future outcomes, such as covid-19 cases, recoveries, and vaccination proportions, and consider how time horizon and changing conditions affect validity.

Discover prescriptive analytics and how what-if analyses turn predictions into actions, as you explore descriptive, diagnostic, predictive, and prescriptive stages with real-world health and manufacturing examples.

Master CRISP-ML(Q), the cross-industry standard process for data science, covering business and data understanding, data preparation, model building, evaluation, deployment, monitoring, and maintenance.

Define the scope and objective to minimize loan defaulters, then train a model with inputs x and output y to predict default and apply survival analytics under business constraints.

Explore data understanding concepts, including data types, scales of measurement, essential terminology, and primary and secondary data collection techniques for supervised learning regression models.

Compare continuous and discrete data, learn how decimals define continuity, and distinguish numeric from categorical data, with examples like time, money, height, weight, and count data.

Differentiate categorical data from count data with binary and multiple category examples, including churn and default scenarios, and outline nominal, ordinal, interval, and ratio classifications.

Explore practical data understanding through real-time examples, identifying nominal, ordinal, interval, and ratio data with travel scenarios, temperatures, and prices.

Explore scale of measurement across nominal, ordinal, interval, and ratio data, highlighting counts, frequencies, percentages, and why ratio data enables broad statistical analysis for data science.

Compare quantitative and qualitative data, distinguish continuous, count, and categorical data, and recognize structured versus unstructured data to inform decision making in regression modeling.

We compare big data and non big data using the three v's and discuss storage and compute needs for structured versus unstructured data, with SQL and NoSQL options.

Explore data collection, distinguishing primary and secondary data sources, and differentiate output variables (response, dependent) from input variables (explanatory, predictors) in structured data for machine learning.

Explore secondary data sources, distinguish primary from secondary data, and learn to blend internal data with Google Maps and drone analytics to improve data-driven telecom insights.

Learn how design of experiments guides data collection for marketing trials, testing discount levels, expiry, and customer radius to optimize coupon redemption.

Identify and mitigate errors in data collection, including random and systematic errors, faulty measuring devices, and measurement procedures to ensure unbiased, representative data for regression models.

Understand bias and fairness in supervised learning, ensure models yield fair results, and emphasize data collection with diverse data, avoiding race or gender as predictors in loan default prediction models.

Explore the crisp-ml(q) data preparation framework, outlining six phases, with phase one on business and data understanding, data types, data collection (secondary to primary), and errors.

Explore the power of probability, from understanding probability distributions to modeling a random variable like daily iPad sales, using discrete versus continuous data and table or graph representations.

Explore the normal distribution as a continuous probability distribution, using heights or profits to illustrate its shape, area under the curve equals one, and zero probability for a single value.

Explore inferential statistics by drawing inferences about a population from a sample using simple random sampling and sampling frames. Learn about hypothesis testing and compare parametric and nonparametric approaches.

Understand the standard normal distribution and z-scores, including how mean and standard deviation shape the curve. Learn standardization with z = (x - mean) / standard deviation and sigma rules.

Explore how box plots relate percentiles, quantiles, and quartiles, defining q1–q4 and their connections to min, median, and max with practical examples.

Use the normal q-q plot to assess normality by comparing sample quantiles with theoretical quantiles and standardized values; if points lie on a straight line, data are normally distributed.

Understand how a bivariate scatter plot reveals the direction and strength of a relationship between two numeric variables, using R to gauge linear, nonlinear, and exponential patterns, outliers, and clusters.

Download Python from python.org, install version 3.10.7 on any OS, note that Python is open source and free for individuals and organizations, and consider Anaconda for a nicer interface.

Learn how to download and install the Anaconda distribution across Windows, macOS, Linux, and Unix, with pre-installed libraries and OS independence that save data scientists setup time.

Explore Anaconda Navigator, Spyder, and popular Python libraries (numpy, pandas, scikit-learn, matplotlib) for reading CSV data, executing code, and visualizing results in a practical learning workflow.

Learn data cleansing and typecasting with Python and pandas, using read_csv and astype to prepare datasets for regression models by converting columns and validating data types.

Learn how to identify and handle duplicates using master data management and data quality concepts, including consolidating records, removing duplicate rows and columns, and recognizing high correlation.

Learn how to identify and remove duplicate records in a dataset using Python and pandas, and explore keep parameters like first, last, and false to drop duplicates and clean data.