Latest Insights
Exploring the frontiers of AI, software architecture, and engineering.
Time Zones, DST and Date Formats: Datetime Hygiene for Analysts
The campaign launched at midnight — in which time zone? Mixed zones, DST gaps and 01/02/03 ambiguity corrupt analysis silently. Rules for datetime hygiene.
Read Article →Why Your CSV Opens Wrong: Encodings, BOMs and Delimiter Guessing
"Müller" renders as "Müller", the first column is named "id", and everything sits in one column. Three CSV gremlins, three fixes — none of them "retype it".
Read Article →Joins That Silently Multiply Rows: Merge Pitfalls in Practice
Revenue doubled overnight — in the report, not in the bank. A duplicate key in a dimension table fanned out every joined row. How to audit merges before they lie.
Read Article →Beyond Deleting Outliers: Winsorization and Honest Treatment
The fastest way to "fix" inconvenient data is deleting it. Winsorization caps extremes instead of erasing them — influence limited, sample intact, method disclosed.
Read Article →Web Workers for Heavy Data Tasks: Keeping the UI Alive
Your CSV parser works — it just locks the page for nine seconds. Web Workers move heavy data tasks off the UI thread. Patterns, transferables, and honest fallbacks.
Read Article →IndexedDB Patterns for Client-Side Data Apps
LocalStorage caps at 5MB and blocks the thread. IndexedDB holds real datasets — if you survive its API. Versioning, bulk writes, indexes, and quota discipline.
Read Article →