Pandas to Polars: Why Speed Matters for Finding the Truth
You can't ignore the performance advantage of Polars anymore. Moving from Pandas' in-place mutations to Polars' declarative expressions lets you verify truth across much larger datasets.
Next 
Val Data
Data insights, analytics, and technical explorations.
Iām Dominick Ryan, a systems and IT operations engineer. I build dependable cloud foundations, automate administrative overhead through code, and profile compute, security, and lifecycle trade-offs across distributed environments. This site documents systems benchmarks, operational architectures, and technical postmortems from platforms I maintain. There are also some personal opinions and musings, which you are free to ignore, embrace, or rail against as suits you.
My Posts
PEP 686: Python Finally Kills the Windows Locale Default
Python 3.15 makes UTF-8 the default across all platforms. Legacy PowerShell 5.1 can still get in the way.
Air-Gap Your AI for Fun and Safety: When Local Models Make the Most Sense
How a hybrid workflow pairing basic Python pre-filtering with a local, air-gapped AI model turned an unmanageable 40 MB production log dump into an actionable root-cause analysis.
Interactive Projects & Tools
Live Transit & Weather
BMT Brighton Line Tracker
Real-time MTA arrival tracking, live weather commute alerts, and historical performance diagnostics for the Q and B corridor.
Launch Tracker →
Macroeconomic Modeling
Housing Market Simulation
Interactive computational model simulating regional listing volumes, inventory absorption, foreclosure rates, and market pricing.
Run Simulation →
Physics & Complexity
Boids Flocking
Observe emergent complexity through three simple steering rules: Separation, Alignment, and Cohesion in a real-time vector canvas.
Launch Simulation →My Reading List
Data Analysis / Python doi.org
Data Structures for Statistical Computing in Python
The classic paper introducing the pandas library to the scientific Python community, detailing the design of high-performance DataFrame data structures for analysis.
Databases infolab.stanford.edu
The Anatomy of a Large-Scale Hypertextual Web Search Engine
The original paper by Google's founders describing the architecture of their prototype search engine, introducing the PageRank algorithm and early web indexing structures.
AI arxiv.org
Attention Is All You Need
The groundbreaking paper that introduced the Transformer architecture, replacing recurrent and convolutional networks with self-attention and forming the foundation of modern Large Language Models.