Half Trillion Location Records Analysed
As Location Intelligence Lead from 2018–2020 designed and constructed the new location intelligence platform for Qrious using over half a trillion records of mobile phone data.
In addition to system design, construction, and documentation, created the trip logic and other algorithms, wrote all analytical code, and designed migration onto cloud.
Costs reduced by an order of magnitude and custom analyses became profitable.
SOFA Statistics(300K downloads)
Sole creator of open source desktop statistics application SOFA Statistics — 300,000 downloads to date. Won People's Choice Award in 2012 NZ Open Source Awards and was a finalist in the Software Project award.
Managing National Data Collection
Responsible for managing New Zealand's gambling treatment data systems for seventeen years
Designed and built all information systems from end-to-end including data collection and reporting.
Active in Tech Community
- Founding committee member of the NZ Python User Group Incorporated. Secretary of the Society for nine years.
- Guest lectured at the University of Canterbury on Statistics for the Digital Humanities course
- Delivered advanced Python training to groups such as engineers from Martin Jetpack while working for open source specialists Catalyst IT
- Guest lectured for NatColl on PHP web development
- Presented Practical Python Async for Dummies to Kiwi PyCon. > 15K views. Also delivered Python and Creative Data Analysis
- 1.5 hour presentation to Spark Data Scientists in 2020 of Pandas for Dummies
Statistical Analysisin Mental Health
Conducted statistical analyses for the Waitemata DHB Clinical Research and Resource Centre (CRRC) for eight years using SPSS.
Also developed the OMA (Outpatient Medication Audit) research database, used in NZ and the UK.
Technical skills
- Strong Python skills
- pySpark, Pandas, Matplotlib, Folium, Mapbox, Jupyter, Sphinx
- Athena, pyAthena, EMR clusters, Glue, S3, boto (AWS); python3 runbooks, Docker on Linux VMs (Azure)
- SQL: Snowflake, Hive, Greenplum, PostgreSQL, MS SQL Server, MySQL, SQLite, MS Access
- NoSQL: Neo4J Graph database (Word Puzzles with Neo4j and Py2neo, redis)
- Django, Flask, CSS, HTML, nginx, gunicorn, javascript, PHP
- Solid Linux skills, git (gitlab, bitbucket, github)
- API integration: Cin7, Dynamics, Xero, Spokephone
Costs and Clinical Inputsfor Seven DHBs
Created AURDA (Analysis and Understanding of Resource Data Application) Reporting Tool to enable seven DHBs (District Health Boards) to better understand their costs and client outcomes.
The AURDA Reporting Project arose out of the national Mental Health Classification And Outcomes Study (NZ-CAOS). The CAOS study was highly technical and complicated, and it was agreed that an easy-to-use, interactive, and graphical reporting tool would enable the participating DHBs to better understand costs and client outcomes.
Academic Research and Presentations
- PhD in Social Sciences — University of Auckland
- John Dobson prize for best opioid research paper, and John O'Hagan prize (twice) for best research presentation at ALAC's Cutting Edge conference
- Struggling Rationalities — Health Information and the Management and Planning of Service Delivery.Two-hour lecture delivered to Sociology Stage III paper (University of Auckland) for 3 years — 82.333 The Sociology of Health and Medicine.
- Socially obligatory drinking: a sociological analysis of norms governing minimum drinking levels, Contemporary Drug Problems
- Social Interactions Involving Alcohol and the Production, Reproduction & Transformation of Interpretations Made of Drinking and Intoxication — Scope for a Symbolic Interactionist Analysis, The Drinking and Drug Practices Surveyor
- and many more national and international publications ...