From policy questions to practical prototypes

In our first publication on the Data & Design Hackathon 2026, we introduced two challenges set by the Yorkshire & Humber Office for Data Analytics (YHODA): where greener homes are most needed in Sheffield, and how neighbourhood change can be tracked without reducing places to static labels. This follow-up shares the work produced by six multidisciplinary student teams at the University of Sheffield and highlights two contributions from each.

Five teams explored greener homes. They combined Energy Performance Certificate (EPC) records with socioeconomic data in different ways, testing housing archetypes, priority scores, maps, dashboards and household-facing tools.

Challenge 1: Where are greener homes most needed in Sheffield?

Starting from a policy question that did not produce one single answer. Each team made different choices about how to combine energy performance, housing characteristics, deprivation, income, retrofit potential and spatial concentration. Those choices are themselves an important policy lesson: a priority score is designed, not discovered, and its results should be tested against secondary sources.

Sheffield Watt Watchers

Team: Thea Uttley (BSc Mathematics); Chrisvyn Kanatazan (BSc Data Science); Sanchaya Sathishkumar Deepa (BSc Data Science); Britiparni Biswas (BA Architecture); and Cindy Ho (BA Architecture).
Insights:

• The team combined EPC data with the 2021 Census, the English Indices of Multiple Deprivation 2025 among other socioeconomic datasets. Using a K-means modelling, initially produced 10 clusters. The team classified two as outliers and used the remaining 8 clusters as exploratory housing archetypes with different energy, carbon, income, deprivation and tenure profiles.
• They designed tools for both sides of retrofit decision-making: a government-facing map and dashboard for exploring priority areas, and a household-facing recommender app connecting each archetype to possible building-fabric, heating, lighting, solar and district energy measures. The prototype also kept affordability, energy use and carbon impacts visible.

Student materials:
- Retrofit Recommender.
- Sheffield Watt Watchers Dashboard for Government recommendations.
- Sheffield Watt Watchers poster.
- Sheffield Watt Watchers presentation.  

Greener Sheffield

Team: Alvia Zehra (BSc Data Science); Zahra Alkhalifah (BSc Data Science); Alzahraa Alessa (BSc Data Science); and Fatemah Al Dhahi (BEng Computer Systems Engineering).

Insights:

• Greener Sheffield created five housing archetypes and a prototype Retrofit Priority Index. The team assigned 40% of the index to energy need, 40% to deprivation and 20% to retrofit potential. Its archetypes ranged from a “Drafty Historic Semi” to a “Modern Pre-Owned Flat”, with a suggested level of intervention for each.
• Using this index, the team highlighted 5 areas in Sheffield in need of energy efficiency upgrading. It also reported statistical associations between energy ratings and selected window and wall characteristics. The results were brought together in an interactive dashboard designed to support further local-authority analysis.

Student materials:
- Greener Sheffield Dashboard.
- Greener Sheffield poster.
- Greener Sheffield presentation.

Tangerine

Team: Jhoshna Udayakumar (BEng Mechatronics and Robotic Engineering); Meena Ramamoorthy (MEng Mechatronics and Robotic Engineering); Himanshu Kriplani (BEng Electronics and Electrical Engineering); Tamanna Mishra (BSc Computer Science and AI); and Dat Chi Le (MPhys Theoretical Physics).

Insights

• Tangerine foregrounded a fairness question: the home with the lowest EPC rating is not automatically occupied by the household least able to pay for improvements. Its approach combined EPC data with the Index of Multiple Deprivation (IMD), income and employment measures. The team then used bivariate mapping and housing-archetype clustering to identify where poor building performance and socioeconomic vulnerability overlap.
• Its dashboard can switch between combined, EPC-only and socioeconomic layers before moving from an area view to property-level detail. For future development, the team proposed a feedback cycle in which completed improvements and updated EPC information would be returned to the dashboard, allowing the effects of funding decisions to be reviewed over time.

Student materials:
- Tangerine poster.
- Tangerine presentation.

SEA The Future

Team: Muhammad Faiz bin Azizan (BSc Data Science); Pyae Hein Kyaw (BSc Economics and Mathematics); Muhammad Irfan bin Khaidir (MEng Electrical and Electronic Engineering); Zul Fauzan bin Zaidi (BSc Data Science); and Thanh Hang Nguyen (MA Education and Psychology).

Insights

• SEA The Future developed an interpretable prototype Priority Score. It assigned 50% of the score to housing inefficiency, 30% to deprivation and 20% to area risk. Its house-level efficiency measure was based on building fabric at 70%, glazing at 20% and floor area at 10%. Properties were also organised into nine housing archetypes, with an additional unclassified category.
• The team used postcode-level summaries to explore concentrations of lower-efficiency housing and the relationship between energy performance and deprivation. It also proposed a dashboard and predictive model for estimating improvement potential and cost-effectiveness. 

Interactive output: Watch the dashboard demonstration

Student materials:
- SEA The Future poster.
- SEA The Future presentation.

bug biriyanis

Team: Ayesha Zia (BSc Data Science); Prachiti Mane (MSc Data Science); Honey Motwani (MSc Data Science); Milan Raje Pradeep Raje (MSc Data Science) and Shipra Jaiswal (MSc Data Science).

Insights

• The presentation reports an analysis of 101,060 EPC records. The team treated records in EPC bands C and D, 78.6% of the dataset, as its largest addressable retrofit segment. 
• The team proposed a dashboard that would show priority areas for retrofit and provide time-series projections through 2030. These projections were presented as a prototype capability rather than as validated forecasts.

Student materials:
- bug biriyanis poster.
- bug biriyanis presentation.

Challenge 2: How do neighbourhoods change over time in Sheffield?

The second challenge asked students to focus on time series analysis rather than labelling places from a single snapshot. The purpose was to identify early signs of decline, renewal and opportunity while avoiding a simple ranking of “good” and “bad” neighbourhoods.

GeoForgers

Team: Taiga Roberts (MSc Human and Molecular Genetics); Rakhi Giri (MSc Data Science); Abishek Ebenezar Ravi (MSc Human and Molecular Genetics); Nana Kwabena Archampong Poku (MSc Human and Molecular Genetics); Gitanjali Yadav (MSc Data Science); Hania Batool (MSc Geographical Information System); and Priyanshi Kataria (MSc Geographical Information System).

Insights

• GeoForgers analysed five indicators: unemployment, population, general health and school qualifications; using 2011 and 2021 Census information alongside open police-recorded crime data. Clustering in R and spatial analysis in ArcGIS Pro led the team to report greater variation between neighbourhood conditions in 2021 and a marked east-west pattern of inequality.
• In an exploratory projection, the team identified parts of central, eastern and south-eastern Sheffield as areas for further investigation and local public investment. It led the team to recommend early, place-based approaches combining housing, employment and community support with continued local monitoring.

Student materials:
- Geoforgers poster.
- Geoforgers presentation.

YHODA thanks all six teams for the ambition, care and range of skills they brought to the challenges, and for showing how multidisciplinary student work can open up new ways to analyse and communicate local data.

These findings, rankings and projections are exploratory student outputs and have not been independently validated or endorsed by YHODA. Differences in data coverage, geographic boundaries and modelling choices mean that further analysis would be required before they could inform decisions about individual households or neighbourhoods.