Open Home Foundation publishes anonymized survey data from 8,616 Home Assistant users
On August 26, 2026, the Open Home Foundation released the anonymized dataset of its first community survey, gathered from 8,616 Home Assistant users, on Zenodo. The lesson for any self-hosted project: you can do open research without exposing your users, if you anonymize properly.
August 26, 2026. The Open Home Foundation publishes the anonymized dataset of its first Home Assistant community survey on Zenodo: 8,616 responses collected between December 2024 and February 2025, now freely available under a DOI. This is not window dressing: the organization ran a k-anonymity review, a GDPR check, and dropped 12% of responses as a precaution before publishing. For the self-hosting ecosystem, it is a case study — you can publish real data about your users without exposing them.
Who actually lives in a self-hosted smart home
The survey, launched on December 16, 2024 and closed in early February 2025, answers a simple question: who installs and maintains Home Assistant? Since the foundation’s creation in 2024, the platform has become the reference for open source smart-home automation, built around the idea of an “open home” that escapes the silos of big tech. Until now, no one knew precisely who was running it. The responses paint an unambiguous profile.
The three reasons for adopting the platform are, in order: local control (89.3%), its open source nature (84.0%), and customization (81.7%). It is no accident that these three come out on top: they are the core of the Home Assistant contract, and the reason a user chooses to self-host their home rather than buy a proprietary hub.
The typical user manages between 11 and 200 devices, lives 75% of the time in a home they own, often across two or more floors. 78.3% of primary users work or study in technology or engineering, and 15.7% identify as neurodivergent. More than half of surveyed households have a cat or dog — a small detail that says a lot about the nature of the data collected, and about what had to be anonymized.
The most revealing figure concerns the division of household labor. In the surveyed homes, it is a single person — 94.4% of them men — who sets up and maintains the automation, while partners (87.4% women) and children interact with the devices without ever touching Home Assistant. One person configures; everyone else lives with the result.
Anonymize first, publish second
The real novelty is not the numbers but the method. Making data “technically available” online is not the same as sharing it fairly. The Open Home Foundation followed the FAIR framework — Findable, Accessible, Interoperable, Reusable — and layered a multi-step anonymization protocol on top.
Strip the identifiers. Names, technical identifiers, timestamps, free-text comments that could hold personal details, and any sensitive information (health, disability, religion, other household members) were removed. So much was deleted that the shared dataset is no longer considered “personal data” under the GDPR.
Run a k-anonymity review. Removing obvious identifiers is not enough: a rare combination of ordinary details can still single someone out. k-anonymity guarantees that every respondent shares their characteristics with at least k others, so that no one can be isolated. The team bucketed over-specific answers — for example “5–6 protocols” instead of the exact list of protocols in use.
Cut rather than risk. When, even so, some combinations remained too rare, the responses were deleted: about 12% of the total, one in eight. The stated rule — publish less and be certain — stands in sharp contrast to how many vendors handle user data.
All of it was legally reviewed and documented: the anonymization protocol, the analysis scripts, and the descriptive statistics are published on GitHub so anyone can audit them. That transparency — publishing the method as much as the data — is what makes the approach credible and reproducible.
What it changes for Home Assistant — and the ecosystem
The data is not sitting in a drawer. It has already shaped the redesign of areas in Home Assistant, the design of the Home Assistant Connect ZBT-2, and the foundation’s privacy research. Publishing also serves open science: other researchers can build on the data instead of re-collecting it, and verify the conclusions rather than taking them on faith.
For self-hosters, the example has immediate practical value. How many community projects publish usage statistics without any protocol, or stay silent out of GDPR fear? The Open Home Foundation shows a middle path exists: collect with consent, anonymize seriously, document the method, and publish under an open license. It is a model transferable to any self-hosted piece of software — a forge, a feed reader, a photo manager — provided you accept the time cost of real anonymization.
The limitations are acknowledged: the survey averaged over 40 minutes instead of the promised 20, and it was English-only. Respondents therefore represent the committed core of the community, not all users. That is a documented bias, not a hidden flaw — and it is exactly what distinguishes a scientific exercise from a PR operation.
The context: the open home against the silos
The survey is part of a larger strategy. Since the Open Home Foundation was created, Home Assistant has stopped being just a piece of software: it aims to be the infrastructure of an “open home” that guarantees data sovereignty against proprietary ecosystems. The foundation is also pushing a community device database, a kind of “Wikipedia of smart-home devices,” fed by real usage data from willing users. Publishing this survey is the first visible step in a more ambitious project: producing open knowledge about the smart home, in the same way the code itself is open.
That ambition has a methodological cost. To build a reliable device database without betraying users, you need to know how to collect, anonymize, and publish cleanly. The December 2024 survey therefore doubles as a proving ground: it validates that the foundation can do what it claims to do, before applying the same rigor to usage data that is far more sensitive than answers to a questionnaire.
The numbers themselves carry a market signal. When 89.3% of users cite local control as the top reason and 84% cite open source, it confirms that self-hosters are not choosing Home Assistant for cost — the free software is often more expensive in time — but for sovereignty. Any vendor or project that underestimates that motivation misunderstands the audience the survey is actually describing.
As for what comes next, the foundation says it plainly: a new survey is in preparation, with questions refined by community feedback — especially around friction points and future use cases. The first edition will at least have established the method; the second will have to reproduce it at a larger scale.
Verdict
If you maintain a self-hosted project, read their anonymization protocol before publishing any user data: the k-anonymity review and the decision to cut 12% of responses are reproducible, and they protect your users as much as your reputation.
If you are a researcher or simply curious, the dataset is a rare look at the reality of self-hosted smart homes — who installs them, why, and who carries the invisible maintenance load. Download it, but remember: it was engineered so that no individual can be recognized, and that is exactly why it is usable.