About

A working notebook on data analytics, modeling, governance and applied AI. The premise is simple: a tool is easy to learn once you understand the problem it was built for, and nearly impossible to use well before that.

Everything here is free to read, with no account and no paywall. Written by a data practitioner working in analytics, data engineering and governance — corrections are welcome and get published.

How this is written

01Idea before toolEvery piece states the problem in plain words before naming a technology. Tools change; the problem is older than all of them.

02Worked examples, not slogansClaims come with the schema, the query or the arithmetic that makes them checkable.

03Failure modes namedWhat breaks, and how you notice — the part most write-ups leave out.

04Sourced and revisableClaims cite their origin and corrections get published rather than quietly patched.

"Learning data isn't about memorising syntax or chasing trends. It's about building mental models that survive the next tooling cycle."

Browse subjects →Latest articles

Reading behind the subjects

Codd, E. F. (1970). A Relational Model of Data for Large Shared Data Banks. CACM 13(6).

Foundations

Kimball, R. & Ross, M. The Data Warehouse Toolkit, 3rd ed.

Data systems & platform

Tukey, J. W. (1977). Exploratory Data Analysis.

Analytics in practice

Wickham, H. (2014). Tidy Data. Journal of Statistical Software 59(10).

Analytics in practice

Sculley, D. et al. (2015). Hidden Technical Debt in Machine Learning Systems. NeurIPS.

AI & applied intelligence

Articles cite their sources; nothing here asks you to take a claim on trust.

About

Everything here is free to read, with no account and no paywall. Written by a data practitioner working in analytics, data engineering and governance — corrections are welcome and get published.

How this is written

01Idea before toolEvery piece states the problem in plain words before naming a technology. Tools change; the problem is older than all of them.

02Worked examples, not slogansClaims come with the schema, the query or the arithmetic that makes them checkable.

03Failure modes namedWhat breaks, and how you notice — the part most write-ups leave out.

04Sourced and revisableClaims cite their origin and corrections get published rather than quietly patched.

"Learning data isn't about memorising syntax or chasing trends. It's about building mental models that survive the next tooling cycle."

Browse subjects →Latest articles

Reading behind the subjects

Codd, E. F. (1970). A Relational Model of Data for Large Shared Data Banks. CACM 13(6).

Foundations

Kimball, R. & Ross, M. The Data Warehouse Toolkit, 3rd ed.

Data systems & platform

Tukey, J. W. (1977). Exploratory Data Analysis.

Analytics in practice

Wickham, H. (2014). Tidy Data. Journal of Statistical Software 59(10).

Analytics in practice

Sculley, D. et al. (2015). Hidden Technical Debt in Machine Learning Systems. NeurIPS.

AI & applied intelligence

Articles cite their sources; nothing here asks you to take a claim on trust.