Machine Learning for Absolute Beginners - Level 3

Explore two popular data visualization libraries, Matplotlib and Seaborn, highlighting their complementary strengths, rich documentation, and how to combine them for quick charts before diving into Matplotlib.

Learn how to create and customize charts in Matplotlib using plt, plotting simple line charts with axis labels, titles, and legends, and explore styling options and axis control.

Understand how Matplotlib constructs charts with figures as canvases and axes as plots, arrange axes in rows and columns, and interact with figure and axis objects.

Matplotlib offers two interfaces, the pyplot and the object oriented interface, for creating charts; the lecture recommends the object oriented approach for flexibility and Python alignment, especially when using seaborn.

Explore the Matplotlib online documentation at matplotlib.org, focusing on API overview, Pyplot and object oriented interfaces, and learn to create figures and axes with subplots and keyword-based customization.

Explore Seaborn's figure and axis objects built on Matplotlib. Learn figure level and axis level functions, using cut plot and swarm plot to create and customize charts.

Master seaborn chart customization with matplotlib, adding titles via axis objects, using sns.set and set_style for themes like dark grid and ticks, and applying hue in scatter plots.

Review Seaborn API reference and gallery demos to explore figure and access level plots, including bar, scatter, and cat plots, with dataset loading via load_dataset.

Explore NumPy, a Python library for large multidimensional arrays and high level mathematical functions, including array creation, shapes, axes, range, line space function, and random data for training.

Learn to select the right chart from Matplotlib and Seaborn, balancing simplicity and insight to clearly convey trends, distributions, and correlations across ranking and proportion charts.

Explore ranking and proportion charts in data visualization for machine learning, including bar variants (grouped, lollipop, stacked), pie charts, treemaps, color optimization, and a final practice with solutions.

Learn to create bar charts that relate a numeric value to a categorical variable, using the Titanic and tips datasets, with matplotlib and seaborn, including horizontal bars and grouped options.

Learn to build grouped bar charts for two categorical variables using seaborn and matplotlib, using estimator sum to show survival by sex and class and revenue by year and product.

Explore how a lollipop chart replaces bars with a line and a marker to visualize sold units by product, including data frame creation, stem plots, and sorting for insights.

Visualize proportions with a pie chart of Group A, Group B, and Group C, using colors and a legend, showing percentages and exploding a slice via Matplotlib.

Master treemap charts that use nested rectangles to show proportional data, including county results colored red to blue by party and voter share.

Learn to optimize colors in charts by using manual color lists or predefined color maps, and apply the cmap parameter with examples like inferno or hot.

Explore line, area, and stacked area charts for trends and distributions across two to three numerical variables, and learn histogram, density curve, box-and-whisker, and swarm visualizations.

Plot time-based data with line charts using Seaborn and Matplotlib, creating multiple lines from value one and value two, and learn about aggregation, means, and confidence intervals.

Explore how an area chart fills space between line and x axis using fill_between to show the evolution of a variable, and shade areas between lines or against zero baseline.

Explore creating stacked area charts with matplotlib’s stackplot, stacking three lines on a baseline to show total revenue and each product’s contribution, while noting readability challenges.

Explore distribution charts with histograms, learn how bins divide numerical data into ranges, visualize frequency, compare bin sizes, and distinguish histograms from bar charts.

Visualize the underlying probability distribution with a density chart and kernel density estimation (KDE) for a smoother view than histograms, and compare male and female age distributions using CD plots.

Compare multiple distributions across groups with box-and-whisker charts, showing medians, variation, and outliers in seaborn, illustrated with iris and tips datasets and the violin chart as an alternative.

Discover the bee swarm chart, an alternative to box charts that presents a one-dimensional, offset scatter-like distribution, highlighting individual data points and outliers across categories.

Discover how to visualize correlation patterns with scatter charts, correlograms, and heatmaps, revealing linear and non-linear relationships and positive or negative correlations.

Analyze scatter charts to reveal relationships between two numeric variables in two-dimensional space, identify correlations, and encode a third variable with color or size.

Explore heat maps, a color-coded visualization of relationships between two variables on axes, and learn to create annotated seaborn heatmaps and correlation matrices from Titanic, iris, and flights data.

Explore how hexbin charts reduce overplotting in large scatter data by aggregating x and y into hexagonal regions colored by density or a third variable.

Recap using Matplotlib and Seaborn to render charts for ranking, proportion, trend, and correlation, then load, clean, and visualize data with Pandas to prepare level four machine learning use cases.