Clearly-explained step-by-step tutorial for implementing transfer learning in image classification — Often we do not have access to a wealth of labeled data or computing power to build image classification deep learning models from scratch. Fortunately, transfer learning empowers us to develop robust image classifiers for our specific classification tasks, even if we have limited resources. In this easy-to-follow walkthrough, we…

Easily access MySQL databases and execute SQL queries in Python — SQL and Python are indispensable tools for data practitioners to work effectively with data. A common use case would be the initial retrieval of data from relational databases using SQL queries, followed by subsequent manipulation and analysis of the data in Python with libraries such as pandas. But did you…

Implementing Amazon's DataWig in Python to impute missing values in tabular data — Missing values in real-world datasets is a common phenomenon that poses a key challenge for all data practitioners. This issue is even more challenging when the dataset contains heterogeneous data types. In this article, we look at how DataWig can help us perform the imputation of missing values in tabular…

Get up to speed with the various data missingness types and methods for imputation — Contents (1) Types of Missing Data (2) Imputation Techniques (3) Python Packages for Imputation (1) Types of Missing Data There are three general types of missing data, best explained with examples. (i) Missing completely at random (MCAR) The likelihood of missing values in a feature is unrelated to any other data features (observed or unobserved). …

Concept and implementation of the global search algorithm to select the best features for machine learning — Feature selection is vital in machine learning as it boosts computational efficiency and predictive performance by keeping only the most relevant predictors. Beyond popular feature selection classes like filter and wrapper methods, global search methods like simulated annealing are powerful techniques at our disposal.

Exploring examples of data science applications across the insurance value chain — The insurance sector is one of the world’s largest industries based on the value of gross premiums, the scale of investments, and its ubiquitous societal role in covering personal and commercial risks. The sheer size of the industry brings a wealth of data and business opportunities, paving the way for…

An easy-to-follow guide to containerizing multi-service ML applications with Docker — Given Docker’s impressive capabilities of building, shipping, and running machine learning (ML) applications reliably, it is no surprise that its adoption has exploded and continues to surge within the data science field. This article explains how to utilize Docker to containerize a multi-service ML application built with H2O AutoML, MLflow…

Final insights (with lecture notes) from the Machine Learning Engineering for Production (MLOps) Course by DeepLearning.AI & Andrew Ng — Realizing the potential of machine learning (ML) in the real world goes beyond model training. By leveraging the best practices of MLOps, teams can better operationalize and manage the end-to-end lifecycles of ML models in a sustainable manner. In this final article of the 4-part MLOps Specialization series, I summarize…

Developing and managing a data science pipeline for detecting fraudulent credit card transactions — Data science promises to generate immense business value across all industries through the compelling capabilities of machine learning. However, a recent report by Gartner revealed that most data science projects fail to progress beyond experimentation despite having sufficient data and intent. To unlock the full potential of data science, machine…