Why reproducible analysis starts with preserving historical data
Operational databases change. Historical snapshots create a stable record for reporting, regulation, model validation, and later review.
Insights
Practical thinking about data, models, and decisions.
Clear explanations of forecasting, data reconciliation, economic modeling, model review, and analytical systems.
Areas of focus
These are the recurring questions that shape the firm’s work and future publications. The emphasis is on methods that make important analysis easier to understand, test, and use.
A forecasting system is more than a statistical model
Data, assumptions, controls, interfaces, scenarios, review steps, and documentation turn a model into a dependable forecasting system.
When reports disagree, start with definitions
Conflicting reports often arise from differences in populations, timing, denominators, exclusions, or source systems.
How small denominator errors can materially change a KPI
Percentages based on different populations cannot always be added or compared directly. Definitions matter.
What litigation and regulatory analysis have in common
Both require careful data handling, visible assumptions, reproducible calculations, clear documentation, and results that can withstand close review.
How to review an inherited model
Trace inputs, check formulas, test assumptions, reproduce outputs, and determine whether the model is suitable for an important decision.
Turning spreadsheet forecasts into controlled analytical systems
Preserve valuable subject-matter knowledge while improving control, repeatability, scenario management, and documentation.
Why AI can accelerate analysis but cannot define the question
AI can speed up coding, debugging, documentation, and research. Analytical judgment still determines what to ask and whether the result makes sense.
A focused first conversation