Blog, page 13

Field notes — one a day since May 2023. Things we shipped, things we broke, and numbers we went and checked.

2026 2025 2024 2023

1,200 posts. , or browse by year and tag below.

Oct 24, 2025

As-of feature computation, in short

Every feature must be computable from information available at prediction time. Write down each one's timestamp, and the leakage disappears.

modelling Machine learning

Oct 23, 2025

Comparing a period to the one before it

period comparison against a real schema, and the clause in it that is doing the work.

sql Analytics engineering

Oct 22, 2025

The time coverage of saas-subscriptions

947 populated days across a 1,063-day span, and what that allows you to compute.

engineering quality Data engineering

Oct 21, 2025

Counting rows is not testing grain

7,000 distinct combinations against 7,000 rows, and what that answers about the table's key.

sql quality warehousing Analytics engineering

Oct 20, 2025

Reading 3 columns instead of 8

Column pruning times partition pruning on a real schema, and the function call that quietly defeats both.

performance warehousing Data platform

Oct 19, 2025

The average fare usd is not the typical one

A 55.1% gap between mean and median on one column, and which of the two answers the question you were asked.

statistics practice Statistics

Oct 18, 2025

Five minutes with city in city-air-quality

5,480 rows, 4 distinct values, and the one fact about city that changes how you query it.

quality practice Data quality

Oct 17, 2025

The top of baseline_score is 1.1× the 99th percentile

A max 1.1× the 99th percentile, and the three different things that could be.

statistics quality Statistics

Oct 16, 2025

humidity_pct by status: a 4% spread

Real group means for one column across 3 segments, and what the pooled average conceals.

statistics practice Analytics practice

Oct 15, 2025

r = 0.10 between sched_dep_hour and arr_delay_min

A measured correlation of 0.10 between two columns, what its square says, and the diagnostic that costs one line.

statistics Statistics

Oct 14, 2025

Correction: Baseline before model

The page claimed more than it should have. What it says now, and where the pattern stops applying.

quality practice Machine learning

Oct 13, 2025

Charting age by arm

Two facts about the columns decide the chart. Cardinality and skew, in that order.

visualization Visualization

Oct 12, 2025

Correlation Explorer works offline

A diverging heatmap of every numeric pair in a dataset, with a click-through scatter so you can check whether the number is telling the truth.

visualization Visualization

Oct 11, 2025

World development indicators, in one paragraph

A tidy panel of thirty countries over twenty-four years with GDP per capita, life expectancy, population, CO₂ and internet penetration — small enough to hold in your head, rich enough for real modelling..

quality practice Data quality

Oct 10, 2025

Reader question — Window functions, properly

The single highest-leverage thing to learn in SQL. Running totals, rankings, period-over-period, deduplication, sessionisation — all of it collapses into one construct.

practice Analytics engineering

Oct 9, 2025