Data Science

What we do with data

• Econometric analysis — regression and causal inference where the question is whether an effect is real and how large.
• Statistical sampling — defensible samples where the full population cannot be examined, and the error bounds that come with them.
• Machine learning methods — applied where they add accuracy, and explained in terms a court or a board can follow.
• Forecasting and modeling — revenue, demand and financial models under stated scenarios.
• Visualisation — turning a large dataset into an exhibit a jury can read.

Where it gets used

• Class certification — whether damages are computable on common proof
• Wage-and-hour analysis across large timekeeping datasets
• Pay equity, hiring and promotion disparity studies
• Market and pricing analysis in antitrust and IP
• Business analytics and strategy outside litigation

In litigation the constraint is not what a model can do — it is whether the method survives challenge. We favor the transparent method over the clever one, and say when the data will not carry the question.

What we need to start

A description of the data — what it covers, how it was collected, and roughly how large. A short look at structure and completeness usually tells us whether the question is answerable before anyone commits to a scope.

Frequently asked questions

Can you tell whether our data supports the claim before we engage?

Usually, from a description and a sample. That short assessment is worth doing first — it is the cheapest step and it changes the strategy most.

Do you work outside litigation?

Yes. Forecasting, analytics and modeling for businesses and agencies are a normal part of the practice.

What are your rates?

We send the rate schedule and a full CV with the proposal after the scoping call.