The Lab — methodology reference

Reference descriptions of the analytical methods I bring to client engagements, with synthetic worked examples.

What this section is

A reference library of the analytical methodologies I bring to client engagements through Mattey Energy services. Each page describes a class of problem, the established methods that address it, when those methods are the right answer and when they are not, and a synthetic worked example to make the approach concrete.

What it is not

These pages are not case studies derived from any operator’s data. No operator-specific dataset, operational figure, or named project appears anywhere in this section. Every numerical example is synthetic and clearly captioned as such.

For the same reason, there is no downloadable dataset on this site. Real engagements run against the client’s own data, under non-disclosure, with all outputs remaining the client’s property.

Why I publish methodology reference pages at all

Three reasons:

1. Specification before sale. A buyer-side reader can read these pages and form a defensible view on whether the methodology fits their problem before the discovery call. That saves time on both sides.
2. Credibility through transparency. The classical formulations are public-domain techniques (Dantzig 1947, Cox 1972, Kaplan–Meier 1958, Ishwaran et al. 2008). My contribution is fluent application to wells engineering problems — these pages make the application visible.
3. Reference for myself. Each page is also a working artefact I use in engagements to anchor scoping conversations.

Underlying toolchain

Python first, by deliberate choice — pandas for the dataframe layer, scikit-learn for supervised baselines, PuLP with the CBC solver for mixed-integer programming, lifelines and scikit-survival for time-to-event modelling, matplotlib for static diagnostics. Visual Studio Code with Jupyter integration as the editor. No proprietary solver, no proprietary BI tool, no closed-source dependency.