Articles

Long-form pieces on how the tools work, where the standard advice breaks down, and what to do instead.

May 20, 2026 4 min read

The data interview questions that actually predict performance

Most data interviews test recall of things that are one search away. Here are the questions that correlate with people being good at the job, and why.

career Analytics practice

Feb 11, 2026 4 min read

dbt, Spark, or just a Python script?

A decision guide for the question every data team argues about, based on data volume, team size, and how much operational burden you can actually carry.

engineering Data engineering

Oct 15, 2025 4 min read

A/B testing when you do not have the traffic

The honest options for a team with two thousand users a week, none of which are "run the test anyway and hope".

experiments statistics Experimentation

Jul 19, 2025 4 min read

The offline–online gap

Your model scored 0.84 in validation and 0.71 in production. Here are the seven causes, ordered by how often each is the actual one.

modelling production Machine learning

Apr 8, 2025 3 min read

Data contracts without the ceremony

The idea is sound and the tooling around it has become theatre. Here is the version that fits on one page and actually prevents incidents.

quality engineering Data quality

Jan 23, 2025 4 min read

How to read a data job posting

What the words actually mean, which requirements are real, and how to tell an organised data team from one that will make you miserable — before you apply.

career Analytics practice

Sep 12, 2024 4 min read

Nine ways a chart lies

A catalogue of the specific distortions that appear in real business charts, most of them accidental, all of them fixable in under a minute.

visualization practice Visualization

May 30, 2024 4 min read

Dimensional modelling is not obsolete

Every few years someone announces that cheap compute has killed the star schema. It has not. Here is what actually changed, and what did not.

modelling warehousing Analytics engineering

Feb 14, 2024 4 min read

The data engineer's guide to timestamps

Timezones, DST, event time versus processing time, and the specific bugs each produces. This is the topic that costs the most engineering hours per unit of conceptual difficulty.

engineering pipelines quality Data engineering

Nov 22, 2023 4 min read

A practical guide to data file formats

CSV, JSON, Parquet, Arrow, and the table formats layered on top. What each is actually for, and the specific ways choosing wrong costs you.

engineering warehousing performance Data platform

Aug 9, 2023 4 min read

Everything you know about p-values, restated carefully

The definition is genuinely awkward, most working explanations are subtly wrong, and the practical consequences of getting it wrong are large. Here is the careful version.

statistics experiments Experimentation

Apr 18, 2023 3 min read

The average is a lie — a field guide to skew

Almost every number a business cares about is right-skewed, and almost every business reports it as a mean. Here is what that does, why nobody notices, and what to publish instead.

statistics practice Statistics