A personal site

Projects, risk, and the people inside them

2015–present

I study the peopleinside complex projects,and build toolsthat take their side.

murmuration · follows your cursor

§ 00 · Statement

For a decade, across five institutions in Australia and Ghana, I have worked on the people side of complex, high-pressure projects: what they ask of the people delivering them, and what holds those people up. My closest interest: mental health and psychosocial risk in project-based work.

Method

Deep learning, neural networks, and text mining, pointed at the data most people overlook: site chatter, social media, survey signal.

Field

Construction and the built environment, but the questions travel: project delivery and cost, public health, energy, and higher education all appear in the record.

Stance

The evidence gets published, then it gets built: PsycheGuard for the people on site, feedforward agents for students, a cost model now used in industry. Research that stops at the paper isn't finished.

Journal articles

1,270

Citations
h-index 16

A$141k

Grants as PI / Co-I

International collaborators

§ 01 · Lines of inquiry

Work

Five threads

Safety and wellbeing on site

How work stays safe and workers stay well: safety culture, behaviour and interventions on construction and high-risk sites, measured with validated indices and instruments. Home of PsycheGuard, his award-winning app for construction workers' psychosocial wellbeing.

2025 Analyses social media conversations to find out how the construction industry talked about and responded to COVID-19 prevention measures.
2021 Reviews the literature to establish how resilience engineering has been conceptualised and measured for organisational safety.
2021 Systematically reviews studies to see how high reliability organisations have been defined and measured in health care safety research.

Award-winning
11 papers · 2016–2025

Delivering projects, cost and value

What makes projects buildable and worth building: tender pricing, value management, procurement, and the competencies behind public-private partnerships and international joint ventures, mostly in developing-country contexts.

2024 Builds and statistically tests a model of the competencies public and private partners each need to reach genuinely win-win infrastructure deals.
2023 Trains deep neural networks on 198 Hong Kong green building projects to predict final cost and duration from early project data, then ships them as a web app.
2022 Reviews prior research to propose the first model linking how joint ventures are controlled to how they perform.

Industry-adopted
11 papers · 2015–2024

Public voice and the classroom

Listening to what people actually say, at scale, to inform policy and teaching: text mining and sentiment analysis of public discourse on energy, higher education and history, plus evaluation of tools for the classroom.

2025 Mines public social media posts to map what Africans are saying about decolonising their universities, using topic modelling and sentiment analysis.
2024 Applies sentiment analysis to social media data to gauge public opinion on retrofitting existing buildings toward net-zero energy performance.
2024 Studies YouTube videos and viewer comments about Ghana's slave castles to see how the public online recalls and interprets the Transatlantic Slave Trade.

In the classroom
5 papers · 2022–2025

Sustainability and the built environment

Modelling the built environment's footprint so it can be lowered: carbon intensity, building emissions, energy demand and the uptake of sustainable construction.

2021 Benchmarks several machine learning models against each other for predicting carbon emissions from buildings.
2020 Compares four machine learning algorithms for forecasting Australia's carbon emissions and ranks them on accuracy and efficiency.
2020 Builds and deploys neural network models to forecast energy demand for Australia, China, France, India and the USA using 1980-2015 quarterly data.

Policy-facing
5 papers · 2017–2021

Foresight for health and environment

Forecasting risks before they arrive: projecting the global burden of hypertension, diabetes and obesity, and predicting how persistent contaminants move through soils, so planners can act early.

2023 Uses time-series forecasting and clustering on WHO data to project hypertension prevalence to 2040 by sex and country.
2021 Measures how the pollutant PFOS binds to 114 Australian and Fijian soils and, for the first time, trains neural networks to predict that binding from soil properties.
2020 Forecasts diabetes and obesity prevalence to 2030 across 183 countries and clusters countries by their combined risk pattern.

Far-reaching
3 papers · 2020–2023

§ · Featured build · award-winning

PsycheGuard

A confidant in your pocket, wherever the work is.

A conversational AI that helps construction workers understand and improve their psychosocial wellbeing: a private, judgement-free first step toward support in an industry where reaching out is hard.

★ Social Impact + People’s Choice Awards · UNSW Digital Makers Competition, 2024

See all four builds →

§ 02 · From the record

Recent papers

5 of 35 outputs · full record & networks →

2025Voices of change: exploring decolonisation in African higher education through social media text miningMugumbate R, Boateng EBJournal of Contemporary African Studies 2025Leveraging Social Media Data to Understand COVID-19 Prevention Measures in Construction: A Machine Learning ApproachBoateng EB, Oteng D, Bonsu DNO, Gopaldasani VBuildings 2024A Triadic Competency Requirement Model for Successful Win-Win Optimization in Public-Private Partnerships: A Structural Equation ModellingEshun BTB, Chan APC, Boateng EB, Oteng DBuildings 2024Public History in Digital Spaces: Public Interpretations of the Transatlantic Slave Trade and Its Implications for History TeachingSosu E, Boadu G, Boateng EB, Appiah-Thompson CInternational Public History 2024Energy Efficiency Retrofitting of Existing Building Stock for Net Zero: Sentiment Analysis of Social media DataTetteh MO, Darko A, Boateng EB, Chan APCRethinking Pathways to a Sustainable Built Environment (Routledge)

→ All 35 outputs · topic map · collaboration network

§ · Borrowed light

All those experiences, even the ones that weren’t the most positive, made me who I am. And I think the diversity of experience makes me very different from everybody else.

Frances H. ArnoldNobel Laureate in Chemistry, 2018