Our blog: useful articles and resources

Keep yourself on the loop and stay updated.

Unlocking Advanced Data Insights with DBSCAN Clustering

The newsletter highlights DBSCAN (Density-Based Spatial Clustering of Applications with Noise), a versatile clustering algorithm that identifies natural clusters of varying shapes and densities in complex data sets. Unlike traditional methods, DBSCAN categorizes data points into core, border, and noise types, allowing for a nuanced understanding of data structures. Its ability to handle noise effectively, adapt to clusters of any shape, and operate without a pre-set number of clusters makes it a robust choice for exploratory data analysis, particularly in scenarios with outliers or irregularly shaped data. The post emphasizes DBSCAN's role in enriching data analysis by uncovering hidden patterns and structures.

Machine Learning

Isomap Revealed: Transforming Data into Insights

"Isomap Revealed: Transforming Data into Insights" is an insightful newsletter focused on the Isomap technique in dimensionality reduction. It highlights Isomap's ability to preserve data structure and unveil complex, multi-dimensional datasets, using the metaphor of unraveling a Swiss roll. The newsletter explains the Isomap process, from constructing neighborhood graphs to revealing hidden patterns by reducing dimensions. It emphasizes Isomap's effectiveness in capturing non-linear relationships, surpassing linear methods like PCA. The piece concludes by showcasing Isomap as a transformative data science tool essential for deciphering intricate data patterns.

Machine Learning

Discover the Power of Dimensionality Reduction: A Tale of Faces and Features

We embark on an insightful journey through dimensionality reduction, spotlighting our adventures with the fetch_olivetti_faces dataset. Our narrative contrasts the capabilities of Principal Component Analysis (PCA) and t-distributed Stochastic Neighbor Embedding (t-SNE) in distilling high-dimensional facial data into just two features. While PCA strives but struggles to differentiate between eight unique subjects, t-SNE triumphs, revealing the intricate details that set each face apart. This exploration not only highlights the strengths and limitations of these techniques but also illustrates their practical implications in data science, especially in areas demanding acute distinction within intricate datasets.

We provide practical, high-quality education in data science, statistics, Python, machine learning, and AI, combining solid theory with hands-on learning.

Connect with us

Created with

Our blog: useful articles and resources

Unlocking Advanced Data Insights with DBSCAN Clustering

Isomap Revealed: Transforming Data into Insights

Discover the Power of Dimensionality Reduction: A Tale of Faces and Features

Featured links

Connect with us