Auckland, New Zealand  ·  36°50′56″S 174°45′36″E

Phil Donovan

Data Architecture  ·  Engineering  ·  Science  ·  AI

Director — Virtus Solutions

Products — eolas.nz

I design data platforms — Snowflake, dbt, Iceberg, AWS — and the models that run on them, for councils, health agencies and insurers. I came to it from geography, after ten years of spatial analysis across public health, transport and urban policy.

Get in touch Download CV — PDF Word

phil@donovan.family  ·  linkedin.com/in/phil-donovan-virtus-solutions  ·  github.com/phildonovan  ·  Auckland, New Zealand

Experience

2013 — present
  1. 2025 — now

    Principal Consultant — Data Architecture, Engineering & AI

    FIRN · Auckland

    • Led the design of Auckland Council's Modern Data Platform, centralising data assets for scale and governance.
    • Built dbt, Snowflake and Fivetran pipelines — including automated ingestion of LINZ land parcels via WFS changesets, keeping council-wide spatial analysis current.
    • Introduced and rolled out AI-assisted development tooling across the engineering team, setting the standards and workflows for its use.
  2. 2025 — now

    Associate

    Principal Economics · Auckland

    • Built the Health Vulnerability and Adaptation dashboard for the Ministry of Health — an R Shiny and Plumber application modelling climate-change health impacts and adaptation options across New Zealand.
    • Applied Treasury CBAx cost-benefit methodology, including declining discount rates, to cost adaptation scenarios down to SA2 level.
    • Engineering on the Cost Intelligence Platform.
  3. 2024 — 2025

    R/SQL Developer, Engineer & Scientist

    AA Insurance · Auckland

    • Developed the pricing emulator that carried the migration from the legacy Protect platform to Duck Creek.
    • Advised on the architecture of their AWS analysis platform, and built custom SageMaker container images so analysts were productive the moment one deployed.
    • Wrote aair, an R package joining Redshift and other sources to the pricing emulator for reproducible pricing simulation.
  4. 2023 — 2024

    Senior Analyst, R Developer & SQL Engineer

    Te Whatu Ora — Health New Zealand

    • Led geospatial advisory for the Queenstown Cryptosporidium outbreak response, mapping contamination risk to direct public health intervention.
    • Built nphsintel, an R package over Snowflake and PostGIS for population health analysis — cut code redundancy by around 75% and trained 10+ analysts on it.
    • Introduced Quarto-based automated reporting and Shiny dashboards for spatial health trends.
  5. 2022 — 2023

    Principal GIS Analyst

    Ministry of Health · Wellington

    • Automated geospatial workflows with the mohanalytics and mohgeospatial R packages, lifting team productivity by around 40%.
    • Built and maintained the ArcGIS Online Portal in Python and R, surfacing interactive layers on health access disparities with ETL keeping them current.
    • Wrote monitoring scripts tracking portal credit and cloud spend, cutting resource overruns by around 30%.
  6. 2018 — 2020

    Spatial Data Scientist

    Geographic Business Solutions · Auckland

    • Deployed and configured the ESRI Enterprise stack — ArcGIS Pro and Server, Survey123, Collector, Operations Dashboard — for client geospatial platforms.
    • Built 3D air pollution models for Seoul using the ArcGIS JavaScript API.
  7. 2016 — 2018

    Spatial & Economic Data Analyst

    MRCagney · Auckland

    • Delivered the National Policy Statement on Urban Development Capacity analysis — regression modelling of the effect of zoning regulation on land prices nationally.
    • Ran the Whangārei residential and business development capacity assessment, building geometric algorithms to encode planning rules.
    • Public transport modelling against the Auckland and Wellington transport models.
  8. 2013 — 2016

    Spatial Data Analyst

    SHORE, Massey University · Auckland

    • Principal spatial analyst on children's urban health research — from cleaning travel-diary data through to accessibility analysis.
    • Co-authored five peer-reviewed papers on children's mobility and the built environment.

Capabilities

Tooling
Languages
R · Python · SQL
Warehouses & engines
Snowflake · Redshift · PostgreSQL · Oracle · SQL Server · DuckDB · Apache Arrow · Iceberg
Pipelines
dbt · Meltano · Fivetran · Singer · PySpark
Cloud & infrastructure
AWS · Docker · Linux · Git
Modelling
Regression · Supervised & unsupervised learning · Spatial models · Cost-benefit (CBAx)
Spatial
PostGIS · ArcGIS · QGIS · FME · GRASS
Apps & reporting
R Shiny · Quarto · Power BI · ArcGIS Portal

Selected work

Open source & projects

Building now

eolas

Official statistics for New Zealand, Australia and the OECD are scattered across dozens of portals, each with its own formats and quirks. eolas puts 700+ datasets behind a single typed API — clean columns, consistent schemas, no merged Excel headers.

I built the whole stack: Singer-spec taps and a Meltano orchestration layer landing data as Iceberg tables on S3 and Glue, exposed through a FastAPI service on AWS, with first-class Python and R clients, a CLI, and a Snowflake share for warehouse users.

Python · R · Singer · Meltano · Iceberg · FastAPI · Snowflake · AWS

nzcensr

An R package that makes New Zealand census data straightforward to import, in both spatial and non-spatial form.

R · Spatial

NZ Economic Overview

An interactive Shiny dashboard over the vs-warehouse API, surfacing New Zealand economic indicators.

R · Shiny

NPS-UDC

National Policy Statement on Urban Development Capacity — quantifying how local government regulation moves land prices across New Zealand's major urban areas, by regression on zoning rules.

Economics · Spatial regression

NDAI-C

The Network Destination Accessibility Index for Children: a tool measuring how reachable everyday destinations are for kids across Auckland, Hamilton, Wellington and Christchurch.

Accessibility · Network analysis

nphsintel

An R package for Te Whatu Ora giving analysts one route into population health data across Oracle and Snowflake, with spatial analysis built in. Cut code redundancy by around 75%.

R · Snowflake · PostGIS

Writing

2023 — 2026

Education

2007 — 2013

Published research

Co-authored
  1. Kids in the City: Children's Use and Experiences of Urban Neighbourhoods in Auckland, New Zealand Carroll, Witten, Kearns & Donovan
  2. Associations between the neighbourhood built environment and out of school physical activity and active travel Oliver, Mavoa, Badland, Parker, Donovan, Kearns, Lin & Witten
  3. Children's Out-of-School Independently Mobile Trips, Active Travel, and Physical Activity Oliver, Parker, Witten, Mavoa, Badland, Donovan, Chaudhury & Kearns
  4. Assessing neighbourhood destination access for children: development of the NDAI-C audit tool Badland, Donovan, Mavoa, Oliver, Chaudhury & Witten
  5. Development of a systems model to visualise the complexity of children's independent mobility Badland, Kearns, Carroll, Oliver, Mavoa, Donovan, Parker, Chaudhury, Lin & Witten

Get in touch

Auckland · NZST

Open to conversations about data platform work, spatial architecture, and anything where the two meet.