Resources
Learning Materials

Resources

Visual carousels, eBooks and downloadable guides covering Python, Data Analytics, Power BI, SQL, Machine Learning, Deep Learning, and Generative AI.

95 resources7 topics

Carousels95

Swipeable visual explainers — read them slide by slide, as a PDF, or as text.

10
Machine Learning

Building with Data: A Machine Learning Journey

Gathering the right data is the foundation of any successful machine learning project. This involves identifying relevant data sources, cleaning messy information, and ensuring data quality. Think of it as preparing ingredients for a recipe – good ingredients lead to a great result. The more comprehensive and reliable your data, the better your model will perform. Data cleaning processes are crucial. This includes handling missing values, removing duplicates, and correcting errors. Data transformation also plays a role, reshaping the data into a format suitable for your chosen machine learning algorithm. Careful data preparation significantly improves model accuracy and reduces bias.

Open carousel
10
Machine Learning

Cleaning Data: Feature Selection for Better Models

What is Feature Selection? Feature selection is a crucial step in machine learning. It involves choosing the most relevant features from your dataset, discarding irrelevant or redundant ones. This simplifies the model, improves performance, and reduces training time. It's like decluttering your toolbox to focus on the essentials for a job.

Open carousel
15
Machine Learning

Decision Trees — How Machines Split Decisions Like Humans

We make decisions every day by asking simple yes/no questions. For instance, deciding what to wear might start with "Is it raining outside?". Based on the answer, you follow a different path of reasoning. Machines can mimic this exact same process using a structure called a decision tree. It's a model that breaks down a complex decision into a series of simpler, binary choices, just like a flowchart.

Open carousel
15
Machine Learning

Gradient Boosting: The Supermodel of Machine Learning

Boosting is an ensemble learning technique. It combines many weak models to create one strong, accurate predictor. Think of it as a team where every member contributes a small piece to the final solution. A weak model, often a shallow decision tree, is only slightly better than random guessing. By itself, its predictions are not very useful. The magic of boosting lies in how it strategically combines these weak links into a strong chain.

Open carousel
10
Machine Learning

Grid Search vs Random Search: Optimizing Models

Grid search is a method for hyperparameter tuning where you define a grid of possible values for each hyperparameter. The algorithm then evaluates all possible combinations within this grid to find the best one. This approach is exhaustive, meaning it checks every possible combination, ensuring no potential best combination is missed. However, this can be computationally expensive, especially with many hyperparameters or a large search space.

Open carousel
10
Machine Learning

Improve Predictions: The Power of Model Stacking

Model stacking combines the strengths of multiple machine learning models. It addresses limitations of individual models by leveraging their diverse approaches to problem-solving. This technique often leads to more robust and accurate predictions than relying on a single model alone. It’s a powerful approach because different models excel at different aspects of prediction. One model might be great at identifying trends, while another is better at handling outliers. Stacking allows us to combine these advantages for improved performance.

Open carousel
10
Machine Learning

K-Means Clustering Explained Simply

Clustering is an unsupervised machine learning technique. Its goal is to group similar data points together. It helps discover hidden patterns in data without prior labels. Think of it as organizing a messy closet. You group shirts, pants, and socks together based on their type. K-Means is a popular algorithm that performs this task automatically.

Open carousel
10
Machine Learning

K-Nearest Neighbors - Learning by Proximity

K-Nearest Neighbors is a simple, versatile machine learning algorithm. It can be used for both classification and regression tasks. Its core principle is that similar data points exist close to each other. It is called an instance-based or lazy learning algorithm. This is because it doesn't build a model during a training phase. Instead, it memorizes the entire training dataset and makes predictions only when needed.

Open carousel
10
Machine Learning

Level Up Your Models: Feature Engineering Power

What is Feature Engineering? It's the art of creating new features from existing data. Think of it as transforming raw data into something more useful for your machine learning models. Good features help the model learn patterns more effectively, leading to better accuracy and performance.

Open carousel
12
Machine Learning

Logistic Regression: It's Not What You Think

Logistic regression has a confusing name. It contains the word "regression," which typically describes predicting a continuous number. But this algorithm is actually used for classification tasks. It predicts which category an observation belongs to. For example, it can classify an email as spam or not spam. It can predict if a transaction is fraudulent or legitimate.

Open carousel
10
Machine Learning

Master the Confusion Matrix

What is a Confusion Matrix? A confusion matrix is a table that helps evaluate the performance of a classification model. It shows how often predictions match actual outcomes. Think of it as a scorecard for your model's accuracy.

Open carousel
10
Machine Learning

Mastering Hyperparameter Tuning

Hyperparameters are settings that control the learning process of machine learning models. Unlike model parameters, they are not learned from data but set before training begins. Examples include learning rate, number of trees in a random forest, or layers in a neural network. Hyperparameters define how a model learns from data. They influence the model's performance and efficiency. Choosing the right hyperparameters is crucial for building effective models.

Open carousel
10
Machine Learning

Mastering Regression Metrics

Mean Squared Error (MSE) measures the average squared difference between predicted and actual values. It gives higher weight to larger errors due to squaring. Lower MSE indicates better model performance. MSE is sensitive to outliers because squaring large errors amplifies their impact. It's always non-negative and in the same units as the target variable squared. This metric is widely used in machine learning for regression tasks.

Open carousel
10
Machine Learning

Naive Bayes: The Simple Genius of Text Classification

Naive Bayes is a family of simple probabilistic classifiers. They are based on applying Bayes' Theorem with a strong "naive" assumption of feature independence. Despite this simplicity, they often work very well, especially for text classification tasks like spam detection. They are called "naive" because it's a bold assumption to think every word in a document is independent of all others. In reality, words often appear together and influence each other's meaning. Yet, this very assumption is what makes the algorithm so computationally efficient and effective.

Open carousel
10
Machine Learning

Random Forests — the power of multiple decision trees

A single decision tree can be a powerful predictor. It asks a series of yes/no questions to classify data or predict values. However, this strength is also its biggest weakness. A single tree often becomes too complex and tailored to the training data. It memorizes the noise and specific details instead of learning the general pattern. This leads to poor performance on new, unseen data, a problem known as overfitting.

Open carousel
10
Machine Learning

Spotting the Unusual: Anomaly Detection Explained

An anomaly is something that deviates significantly from what is considered normal or expected. It's an unusual event, pattern, or observation. Recognizing these deviations is crucial in many fields, from cybersecurity to finance. Identifying anomalies allows us to investigate and potentially take corrective action before problems escalate. Detecting an anomaly doesn’t always require a precise definition of “normal.” Sometimes, the expected behavior is complex and difficult to determine. Statistical methods and machine learning algorithms are employed to identify patterns and flag those that fall outside those patterns.

Open carousel
10
Machine Learning

Support Vector Machines — The Art of Perfect Separation

Imagine you have two distinct groups of data points on a graph. Your goal is to draw a line that best separates them. This is the fundamental task of classification, and Support Vector Machines are masters at it. They don't just draw any separating line. They search for the one that creates the widest possible "street" between the groups. This street is called the margin, and its width is crucial for the model's performance.

Open carousel
12
Generative AI

TOON vs JSON - The Data Format Battle

JSON remains the universal standard for transferring data, but struggles with AI-driven workloads. TOON is built for large language models (LLMs), aiming to shrink data size for faster AI processing.

Open carousel
10
Generative AI

Types of RAG

Naive RAG — The Basic Approach Naive RAG follows the simplest RAG pipeline: retrieve relevant information and give it to the LLM for generation.

Open carousel
10
Machine Learning

Understanding Prediction Performance: ROC & AUC

The Receiver Operating Characteristic curve visualizes how well a classification model distinguishes between different classes. It plots the true positive rate against the false positive rate at various threshold settings. Think of it as a graph showing the trade-off between correctly identifying positive cases and incorrectly identifying negative cases as positive. A higher ROC curve generally indicates a better model; a model that effectively separates positive and negative instances. Different classification algorithms will produce varying ROC curves, representing their different strengths. Understanding ROC helps us compare and select the best model for our specific problem.

Open carousel
10
Machine Learning

Why Feature Scaling Matters

Feature scaling is a technique to adjust the range of features in a dataset. It ensures that all features contribute equally to machine learning models. Without scaling, features with larger values can dominate the model. Normalization and standardization are two common scaling methods. Normalization scales data to a fixed range, usually between 0 and 1. Standardization transforms data to have a mean of 0 and a standard deviation of 1.

Open carousel
10
Machine Learning

Wisdom of the Crowd in ML

What is Ensemble Learning? Ensemble learning combines multiple machine learning models to improve performance. Instead of relying on a single model, it leverages the strengths of many. This approach often leads to better accuracy and robustness.

Open carousel
No matching resources found

Try different keywords or select "All" categories.