Research and Publications

Introduction

This article presents the TAIAO programme, a national research initiative advancing time‑evolving data science and artificial intelligence to support open, scalable, and environmentally focused analytics across Aotearoa New Zealand. By developing new methods, datasets, and software for dynamic, real‑time environmental modelling, TAIAO enables researchers, industry, and government to better understand and

Problems addressed

• Environmental systems are dynamic and rapidly evolving, yet many analytical tools assume static data.

• Existing environmental datasets are often fragmented, inconsistent, or difficult to access, limiting scientific progress.

• Traditional modelling approaches struggle with real‑time data streams, concept drift, and large‑scale environmental variability.

• There is a need for open, reproducible, and collaborative AI tools that can be used across disciplines and sectors.

Methods overview

The TAIAO programme develops new algorithms, data infrastructures, and open‑source tools for analysing time‑evolving environmental data. Its work spans streaming machine learning, scalable data processing, deep learning for environmental sensing, and reproducible scientific workflows. The programme integrates expertise from computer science, ecology, climate science, and industry partners to create a unified platform for environmental AI research in Aotearoa.

Key outcomes

• Establishes a national research programme focused on environmental AI and time‑evolving data science.

• Develops open‑source tools and datasets for real‑time environmental modelling.

• Advances methods for streaming machine learning, concept drift, and large‑scale environmental analytics.

• Supports cross‑sector collaboration between academia, government, and industry.

• Provides a foundation for scalable, reproducible, and transparent environmental science in Aotearoa.

Findings

The article highlights TAIAO’s progress across multiple research streams, including advances in streaming analytics, large‑scale environmental datasets, and AI‑driven modelling tools. Early results demonstrate the programme’s ability to support real‑time environmental monitoring, improve predictive modelling, and enable open, collaborative research through publicly available software and data resources.