Forecasting

Bespoke forecasts for planning and analysis

Forecasts are an essential tool for many organisations. Schiff Consulting creates bespoke forecasting models that allow clients to predict future trends, analyse alternative scenarios, and understand the uncertainty associated with forecasts. This includes applying advanced econometric and machine learning techniques such as time-series cross-validation and ensemble models to produce robust forecasts from limited data.

As well as one-off forecasts, we can develop semi-automated forecasting systems that are updated and re-trained as new data becomes available and allow users to specify their own forecasting scenarios or assumptions to test the impacts of decisions and external shocks.

Examples of work

Government revenue forecasting: The NZ Ministry of Transport commissioned Schiff Consulting to re-develop the model it uses to forecast government revenues from fuel taxes and road user charges. This included developing econometric forecasting models for revenue streams and embedding these models in an R-based forecasting system that can be used to analyse and compare forecast scenarios.

Population and household forecasting: Peer review of models used for local population and household projections, for Taupō District Council.

Aviation forecasting: Econometric modelling to support passenger movements forecasts for Brisbane Airport (with Fresh Info).