Complete Data Science BootCamp

Explore NumPy, the Python library for fast array operations, learn its installation with pip, import as np, create arrays with np.array from Python lists, and understand basic array concepts.

Create numpy arrays with np.array and explore zero-, one-, two-, and three-dimensional structures. Understand array creation, printing, and using ndim to check dimensions.

Explore the shape of arrays and how reshaping changes dimensions and element counts using numpy, including 1D to 2D conversions and flattening.

Learn numpy array indexing to access elements in one- and two-dimensional arrays using zero-based indices. Apply examples like a[0], a[2], and a2d[i, j] to understand element retrieval and basic operations.

Learn how to iterate numpy arrays using Python for loops, traversing 1d and 2d arrays and printing each scalar element with nested loops.

Master NumPy array slicing in Python, using start and end indices and steps, for 1D and 2D arrays, with practical examples and notebook demos.

Learn numpy array searching with np.where and sorting with the sort function, locating value indexes and ordering numeric and alphabetic arrays.

Explore pandas, the Python library for loading data into data frames, cleaning and analyzing datasets. Master the data science lifecycle from CSV input to preparing data for analysis.

Explore pandas series as a one-dimensional data structure, acting like a data frame column and holding any data type, created from lists or dictionaries with pd.Series.

Explore the Pandas DataFrame, a two-dimensional data structure with rows and columns, and learn how to load CSV or Excel data into a DataFrame for data science and ML.

Learn to read csv files with pandas read_csv, store them in a dataframe, and print the data from a csv in the same directory as your Jupyter notebook.

Master pandas data frames analysis by using head and tail to view data frames, info to summarize structure, describe for statistics, and correlation to explore relationships, including read csv examples.

Install and import matplotlib to enable data visualizations, then plot a simple line chart using x and y values with plt.plot and plt.show.

Explore essential Matplotlib plots, including line charts, bar charts, scatterplots, pie charts, and histograms, with practical data, plotting techniques, and line style variations.

Learn to use Seaborn in Python to create box plots, distribution plots, and rich plots for visualizing data. Explore in-built datasets such as tips, Titanic, MPD.

Identify and understand missing values in datasets, then apply practical imputation techniques—mean for numeric, mode for categorical features, using a simple imputer with pandas.

Explore feature encoding techniques for categorical data, including nominal and ordinal methods like one hot, dummy, mean, label, and target guided ordinal encoding, with practical pandas demonstrations.

Explore feature scaling, a preprocessing technique that standardizes features with normalization and standardization. See how min-max and standard scalers adjust features such as age and salary for robust modeling.

Machine learning, a subset of artificial intelligence, learns from data to predict and decide; explore its lifecycle, training data, model building, and supervised, unsupervised, and reinforcement learning.

Explore how supervised machine learning uses labeled training data to train models and map inputs to outputs. Understand the training vs test data lifecycle, and distinguish regression from classification.

Explore unsupervised machine learning, where models use unlabeled data to discover hidden patterns and structure, performing clustering and association, including market basket analysis.

Master the machine learning lifecycle from data gathering and exploration to preprocessing, training, and deployment. Identify data sources, quality, cleaning, and evaluation metrics for classification, regression, clustering, and association.

Use a train test split in supervised learning to train on the 75% training set and evaluate on unseen test data, using X_train and y_train, and X_test and y_test.

Learn regression analysis to model the relationship between a dependent variable and independent predictor variables, enabling predictions of continuous values such as salary or temperature.

Explore linear regression, a predictive analysis method that models the relationship between y and X, and implement it in Python, with a practical SQL example using salary data.

Explore logistic regression, a popular supervised classification algorithm predicting a categorical variable with probabilities via a sigmoid function, and implement it in Python with a train-test split and threshold.

Explore the k nearest neighbors algorithm, a non-parametric, lazy-learning method for classification (and regression) based on Euclidean distance, with a practical Jupyter notebook implementation and scaling.

Explore SVMs for classification, building the optimal hyperplane via margin from support vectors. Implement with a linear kernel on a cat-versus-dog dataset using an 80/20 train-test split and assess accuracy.

Explore how the decision tree algorithm, a classification and regression tree, functions in supervised learning for classification and regression, detailing root and leaf nodes, branching, pruning, and overfitting.

Explore how random forest, an ensemble of decision trees, uses majority voting to boost classification and regression accuracy while preventing overfitting.

Explore k means clustering, an unsupervised learning algorithm that forms k clusters by assigning points to centroids and updating centers, using the elbow method to pick k.

Learn how hyperparameters drive learning and how grid search CV identifies the best C, gamma, and kernel for a support vector classifier, with a breast cancer dataset example.

Explore how a machine learning pipeline automates data ingestion, cleaning, preprocessing, model training, and deployment through modular steps, illustrated with a pipeline using PCA, scaling, and a decision tree.

Learn to evaluate machine learning models on unseen data using confusion matrices, accuracy, precision, and recall, and apply regression metrics like the absolute error and mean squared error.

Learn how cloud computing delivers on-demand data storage and computing over the internet. End users access services from data centers and providers offer SaaS, PaaS, and IaaS.

Discover how Amazon Web Services provides a scalable, cost-effective cloud platform with regions and availability zones, on-demand services, and a user-friendly management console.

Explore AWS services like S3, RDS, IAM, EC2, and SageMaker to understand storage, durability, storage classes, availability, scalable compute, and ML model deployment in the cloud.

Explore how Amazon SageMaker delivers a fully managed ML workflow across the complete machine learning lifecycle—from data labeling and preprocessing to training, tuning, evaluation, and deployment.

Create your first machine learning project on AWS SageMaker notebooks, upload diabetes.csv, train a random forest classifier with pandas, measure ~78% accuracy, and terminate resources to save costs.

Discover built-in AWS SageMaker algorithms for supervised learning, including classification and regression, with pre-trained models like Cat Boost and linear learner.

Implement a linear learner model in AWS SageMaker using S3 data, train on the UK breast cancer diagnostic dataset, deploy a hosted endpoint, and evaluate accuracy.

Explore no-code machine learning with AWS SageMaker Canvas, uploading data to automate data cleaning, processing, and model selection, then build and generate single or bulk predictions.

Explore the AWS SageMaker marketplace to access pre-trained state-of-the-art model packages for tasks like object detection with YOLOv3, review pricing, and try product demos by uploading images.

Explore artificial neural networks inspired by the brain, with dense input, hidden, and output layers, weights and biases, and backpropagation to improve predictions.

Explore how activation functions decide neuron activation, enable non-linear transformations, and bound outputs for regression and classification, featuring sigmoid, tanh, and relu variants.

Explore how optimizers adjust neural network weights and learning rate to minimize loss, covering gradient descent, stochastic and mini-batch methods, and the Adam optimizer's efficiency and stability.

Explore convolutional neural networks and learnable filters that extract hierarchical features from images. Build practical convolutional neural network pipelines with pooling, stride, and TensorFlow and Keras.

Explore recurrent neural networks with feedback loops and memory of inputs. Understand backpropagation through time and how long short term memory networks and gated recurrent neural unit networks mitigate gradients.

explore diabetes prediction using a machine learning workflow: load a diabetes dataset, preprocess with standardization, train a linear svm, evaluate accuracy, and predict a new case.

Predict medical insurance costs with a linear regression model using age, sex, BMI, smoker, and region; encode categorical features, perform a train-test split, and evaluate performance.

build a neural network regression in TensorFlow to predict gold price from spx, oil, silver, and eur/usd, using min-max scaling and two dense layers, achieving about 0.73 r-squared.

Learn to implement a CNN with Keras and TensorFlow, using CNN layers, max pooling, flatten, and dense layers; normalize data, one-hot encode ten-class labels, and train for 20 epochs.

Explore stock price prediction with an lstm-based neural network using Microsoft stock data, including data preprocessing, scaling, and training within a sequential model.