Complete A.I. & Machine Learning, Data Science Bootcamp

Learn how computers use programming and machine learning to solve tasks, from simple rules to complex apps. Explore applications like self-driving cars, vision, language processing, and recommendations.

Train a YouTube recommendation engine by using user likes and dislikes, along with video length and likes, to predict what to recommend.

Machine learning uses data-driven patterns discovered by models to predict future outcomes. It differs from traditional algorithms by learning, and ties data analysis and data science to building these models.

Explore how to define a machine learning problem and choose among supervised, unsupervised, and transfer learning, using classification, regression, clustering, and recommendations to solve real-world tasks.

Explore the types of data in machine learning by distinguishing structured and unstructured data, and static versus streaming data, with practical examples like CSV files and pandas workflows.

Explore evaluation metrics across machine learning tasks, from classification accuracy to precision, recall, mean absolute error, mean squared error, and precision at k, with a real-world 95% accuracy example.

Identify and apply feature engineering to features from structured or unstructured data, including numerical, categorical, and derived variables, to improve heart disease predictions and ensure feature coverage.

Tune machine learning models by adjusting hyperparameters on training or validation data to improve performance, e.g., random forest trees and neural network layers. Compare models on the test data.

Experimentation drives an iterative, tool-matching approach to machine learning projects, guiding you from problem definition through data, evaluation, features, and modeling to refined, hands-on projects.

Choose your path based on Python knowledge: continue with the course or learn Python from scratch, both leading to essential data science and machine learning skills.

Create and manage reproducible data science environments with Mini Conda and Conda, installing tools like pandas and matplotlib for heart disease classification, and share exact environments to avoid dependency issues.

Set up a Windows data science workflow using Mini Conda, create and activate a project environment, and install Jupyter to run notebooks for machine learning and data science projects.

Learn to set up and activate a conda environment, start a Jupyter Notebook, and work with notebooks to import data, run Python code, and create code and markdown cells.

Master keyboard-driven Jupyter notebooks by using shift+enter to run cells, escape and enter for command and edit modes, and pandas to read_csv into a dataframe for plotting.

Learn how to save, run, and share Jupyter notebooks on GitHub for data science and machine learning, activate environments, manage terminals, and collaborate on heart disease projects.

Explore how to describe and summarize data in pandas by inspecting data types, numeric columns, and statistics using describe, info, and column selection on a data frame.

Shuffle a data frame with sample to randomize order for training, validation, and test splits, then reassign, reset index, and apply lambda for converting kilometers to miles.

Learn how to download the pandas exercises notebook from zero to mastery ml GitHub repo using the raw button, save it, and run it in a local environment or Colab.

Explore why NumPy powers data science and machine learning with fast, C-backed operations on arrays, and why converting data to numbers (one or zero) enables machines to learn.

Explore NumPy, the backbone of data science and machine learning, turning data into numerical arrays with vectorization and broadcasting for fast, efficient computations.

Explore NumPy data types and attributes by creating arrays, inspecting shape, dimensions, and dtype, and learn how NumPy underpins pandas dataframe for numerical computing.

Learn to create numpy arrays from lists, generate ones, zeros, and ranges, and produce random arrays while inspecting shapes, data types, and docstrings for numpy concepts and pseudo random numbers.

Explore how to view and manipulate arrays and matrices with NumPy, practice indexing and slicing by shape, and extract unique elements to prepare data for analysis.

Master array manipulation and element-wise arithmetic in numpy, explore broadcasting rules across shapes, and read documentation to resolve shape errors while preparing for aggregation.

Explore aggregations on NumPy arrays by comparing Python and NumPy sums, and apply mean, max, min, variance, and standard deviation along axes to understand data dispersion.

Explore standard deviation and variance as measures of data spread using numpy arrays, comparing high and low variance data with practical array demonstrations, aggregation functions, and histogram visuals.

Explore how to use numpy comparison operators on arrays to produce boolean results for elementwise greater than, greater or equal, and equality checks.

Master sorting arrays with numpy by using np.sort and np.argsort to sort values and indices, and apply arg min and arg max along axes to locate minimum and maximum elements.

Turn images into numpy arrays using image read, explore shapes and color channels, and apply numpy operations to turn data into numbers for machine learning.

Treat imposter syndrome as a natural part of learning, signaling you’re not there yet and that practice builds comfort. Teach others on Discord to reinforce what you’ve learned.

Learn to set up a conda environment, run a Jupyter notebook, and import numpy, pandas, and Matplotlib. Master the object orientated API using plt.subplots to create figures and axes.

Explore the anatomy of a matplotlib figure, from the figure canvas to axes and subplots, including titles, axis labels, legends, and inline plotting in Jupyter.

Explore histograms and subplots to visualize distributions and compare multiple plots, including horizontal bar graphs and a four-subplot figure, in a hands-on code-driven session.

Learn how to create meaningful, accurate data visualizations that honestly convey findings, avoid misleading graphs, and validate insights with careful axis labeling and multiple methods.

Format the price column by removing dollar signs and separators, convert to integers, and compute a cumulative total, then plot total sales over time using pandas data frame plotting.

Practice plotting with pandas by exploring histograms, bin adjustments, and subplots on heart disease data, then compare the stateless pandas plot method with the more flexible object-oriented plotting API.

Explore how to customize plots with matplotlib styles, including seaborn and ggplot, and add titles, axis labels, and legends using the pandas plotting API.

Export your customized plots for presentations by saving as PNG or using the savefig method, enabling shareable visuals from your heart disease data analysis.

Explore the psychic learn machine learning library with hands-on practice and clear explanations. Prepare for milestone projects that reinforce workflows, while getting support through assignments and our Discord community.

Explore scikit-learn workflows from data preparation to fitting, evaluating, and hyperparameter tuning, then saving models for classification and regression using heart disease and Boston data sets.

Explore debugging warnings in Jupyter by using the warnings module, optionally filtering or ignoring alerts, and upgrading conda packages to resolve Python version conflicts.

Convert data to numbers by turning categorical features into one-hot encoded variables with a column transformer, enabling a random forest regression on the car sales dataset.

Master data preparation by imputing missing values with scikit-learn's simple imputer and column transformer, then one-hot encode categorical features and convert to numerical data for model readiness.

Choose the right estimator for regression and classification problems using the machine learning map. Practice building and evaluating models on California housing data with train/test splits and r-squared metrics.

Explore how a decision tree enables machine learning models to program themselves from data, using if-else blocks to predict outcomes without explicit instructions.

Compare linear SVC and random forest on the heart disease classification data to show how ensemble methods boost accuracy for structured data.

Learn to make predictions with regression models using the random forest regressor, including preparing features and labels, splitting data, fitting the model, and evaluating with mean absolute error.

Explore how cross validation goes beyond a single score by using cross_val_score with fivefold splits and a scoring parameter, comparing mean accuracy and coefficient of determination.

Evaluate classification models using accuracy, area under the roc curve, confusion matrix, and classification report. Use a random forest with fivefold cross-validation to estimate cross-validated accuracy on heart disease data.

Explore how a confusion matrix compares true labels to model predictions, visualize it with cross tab and seaborn heatmap, and interpret true positives, true negatives, false positives, and false negatives.

Leverage confusion matrices to evaluate classification models by comparing predictions to true labels. Visualize where the model confuses classes and identify true/false positives and negatives.

Learn how a classification report consolidates precision, recall, F1, accuracy, and macro and weighted averages with the confusion matrix to evaluate a classification model.

Explore regression model evaluation metrics, including the R2 score, mean absolute error, and mean squared error, and learn to score predictions on a train/test split using random forest regression.

Evaluate cyclone models using classification metrics such as accuracy score, precision score, recall score, and F1 score. Include regression metrics like R2 score, mean absolute error, and mean squared error.

Improve a machine learning model by adding data and richer features, and by tuning hyper parameters with methods like manual tuning, random search CV, or grid search CV.

Explore correlation analysis to identify which attributes correlate, with examples like land size and floor space, and learn to use forward and backward attribute selection to improve or simplify models.

Putting it all together demonstrates end-to-end machine learning workflow, from data cleaning and encoding to building and evaluating a pipeline with a random forests regressor to predict car sale price.