Cyberate's research uses operational, public and partner data under explicit boundaries: what the research question needs, what can be analysed, what must be de-identified, and what cannot be used without permission.
Every research question draws on some combination of three source families, each with its own governance.
From live group operations: Generated by DDDI Group's development, construction and supply businesses, and by Cyberate systems in live deployment. This is the operational core most research teams never get. (Development milestones and project status; Scheduling and site records; Procurement events; Quoting workflow data; Reporting dashboards): De-identified before research use.
From open and government sources: The public data layer under every housing decision, assembled and quality-checked before analysis. (Planning records and development applications; Housing market data; Census and demographic data; Infrastructure data; GIS and spatial data): Public does not mean unrestricted: sources and licences are respected.
From partners and fieldwork: Data generated by the research itself, with university partners and industry participants. (Surveys, interviews and workshops; University partner datasets; Model evaluation datasets): Collected under agreed research protocols.
METHOD
Eight steps. Every question.
The same methodology carries a question from definition to something the industry can use, with governance built into the middle, not bolted on at the end.
Define the question: Frame it precisely, and agree what a useful answer looks like.
Identify the sources: Which operational, public and partner data the question actually needs.
Assess data quality: Coverage, gaps and bias, established before analysis begins.
De-identify & govern: Sensitive data de-identified; access limited to the research team on the question.
Analyse & model: Trends, timelines, patterns and the factors behind different outcomes.
Validate live: Findings tested against live operations and practitioner experience.
Document limitations: What the data supports, and what it does not, stated plainly.
Translate: The answer ships as a system, report or dashboard, not just a finding.
The boundaries we don't cross.
Trust in the research depends on how the data behind it is handled. These rules apply to every engagement, commissioned or internal.
De-identification by default: Operational data used in research is de-identified; commercial specifics stay confidential.
Client data stays client data: Customer systems' data is never used for research without explicit agreement.
Agreed purposes only: Data is used for the research purpose it was agreed for. No quiet reuse.
Governed access: Access is limited to the research team working on the question.
No manufactured findings: We publish what the data supports, including the null results.
Limitations disclosed: Coverage, assumptions and methodological limits are stated, not hidden.
Data you can trust the boundaries of.
If your question involves your own operational data, the boundaries are agreed before the research starts. Bring the question; we'll bring the method.