Forecasting: Principles and Practice - The Pythonic Way, The PostgreSQL VScode Extension
This week's agenda:
Open Source of the Week - The PostgreSQL VScode extension
New learning resources - Logistic regression from scratch with Python, architecting agentic RAG systems, Cursor settings, tokens and parameters in LLM, MLOps pipeline with Python
Book of the week - Forecasting: Principles and Practice - The Pythonic Way by Prof. Rob J Hyndman, Prof. George Athanasopoulos, Azul Garza, Cristian Challu, Ph.D. , Max Mergenthaler Canseco, and Kin G. Olivares
I share daily updates on Substack, Facebook, Telegram, WhatsApp, and Viber.
Are you interested in learning how to set up automation using GitHub Actions? If so, please check out my course on LinkedIn Learning:
Open Source of the Week
Microsoft recently released an official extension for working with PostgreSQL on VScode, and it is amazing!
Microsoft has just launched its official PostgreSQL extension for Visual Studio Code. The extension provides a unified database development workflow, integrating schema management, query editing, and AI-powered support—all within the VS Code editor
Core Features
Schema Visualization:
Quickly generate graphical schema diagrams via a right-click in the Object Explorer
AI-Powered GitHub Copilot (@pgsql):
Context-aware IntelliSense and natural-language SQL generation
Special features like “Rewrite Query”, “Explain Query”, and “Analyze Query Performance”
Multi-stage Chat Agent mode: can list servers, draft scripts (e.g., create database, enable extensions), and await user approval before execution
Connection Management
Supports local, Docker, and Azure PostgreSQL
Browse and filter Azure databases
Passwordless authentication via Entra ID, with automatic token refresh
Database Explorer & Object Management
Manage schemas, tables, views, functions, and more—create, modify, or delete these directly in the editor.
Query Editing & History
SQL IntelliSense (keywords, table names, functions), syntax highlighting, and formatting
Session-based query history pane for easy retrieval and reuse
More details are available in the following video from the VScode team:
License: MIT
New Learning Resources
Here are some new learning resources that I came across this week.
Logistic Regression From Scratch in Python
The following tutorial by NeuralNine focuses on the theory behind logistic regression and provides a step-by-step guide for building it from scratch using NumPy.
Architecting Agentic RAG Systems: Transform NPCs Into AI Agents
A new article by Paul Iusztin from the Decoding ML newsletter about architecting agentic RAG systems in production. This includes the following topics:
Architecting scalable and modular RAG feature pipelines
Taking Python to production on AWS
Architecting agentic RAG systems for transforming NPCs into adaptable agents
Making Cursor smarter (and up to date)
Leveraging AI agents relies heavily on providing a good context. The following tutorial by mehdio from the Mehdio's Tech (Data) Corner newsletter provides a step-by-step guide for setting up documentation context and Cursor rules in the Cursor IDE.
Tokens & Parameters in Large Language Models
The following article by Kannan Kalidasan from The Egg Puff Engineer newsletter provides an introduction to tokens in AI models.
MLOps Pipeline with Python, AWS, Docker
A new MLOps course by Boktiar Ahmed Bappy and freeCodeCamp focuses on setting up an end-to-end MLOps pipeline. This includes using tools such as MLflow, Docker, and AWS with Python.
Book of the Week
This week's focus is on a time series analysis and forecasting book. One of the best time series books - Forecasting: Principles and Practice, by Prof. Rob J Hyndman and Prof. George Athanasopoulos, an R book, has a new edition for Python co-authored with Azul Garza , Cristian Challu, Ph.D. , Max Mergenthaler Canseco from Nixtla , and Kin G. Olivares from Amazon.
The book is based on the 3rd edition of the book, and it provides an introduction to the foundation of time series analysis and forecasting with Python. As you can imagine, the code examples are based on the nixtlaverse libraries.
The book covers the following topics:
Fundamentals of time series analysis: visualization, decomposition
Exponential smoothing models and state‑space frameworks
Auto‑regressive Integrated Moving Average (ARIMA) models
Dynamic regression models and causal forecasting
Judgmental forecasting and combining quantitative and qualitative insights
Forecast evaluation: error metrics like MAE, RMSE, MAPE, and model comparison techniques
New chapters - neural networks, and foundation forecasting
The book currently has only an open online version.
The book's ideal audiences are data analysts and data scientists who want to get started with time series and R practitioners who are pivoting to Python.
Have any questions? Please comment below!
See you next Saturday!
Thanks,
Rami
Thanks for the shout out 🥂