Safety here means something broader than cameras and command centres — it covers anything designed to warn people, or a city, before something goes wrong: a disease outbreak, a flood, or the slow loss of a historic neighbourhood. The projects below span all three.
Cauayan's C-DEWS pairs IoT mosquito traps with a neural network that identifies disease-carrying Aedes aegypti eggs at 99.5% accuracy, catching outbreak risk before the first hospital admission. Lilongwe's contribution to Malawi's national health platform works at a different scale entirely: feeding vector-borne disease and cold-chain data into the DHIS2-based One Health Surveillance system behind the country's electronic Integrated Disease Surveillance and Response programme, in use since 2015 to speed up outbreak alerts nationwide.
Rio de Janeiro's Operations Center (COR) is the starkest example in this dataset of a safety system built in direct response to tragedy: opened on 31 December 2010, about eight months after April rains killed nearly 70 people and left 15,000 homeless, it's a $23 million public-private partnership with IBM that today converges 2,500+ camera feeds and data from 60+ agencies into a single room — cutting emergency response times by roughly 30%. Not every early-warning system in this dataset waited for a disaster first, but COR is a reminder that many did.
Havana's digital mapping of Old Havana is a different kind of safety project: OpenStreetMap Cuba volunteers walked the pedestrian streets of the UNESCO-listed historic centre in 2016 with phones and 360-degree cameras, part of the city historian's broader rehabilitation programme — treating the risk of losing an irreplaceable place to neglect as seriously as a flood or an outbreak.
Whether the thing at risk is a person, a neighbourhood or a historic core, all four projects share the same logic: the earlier the warning, the cheaper the response. See the Society category for the wider set of culture, health and inclusion projects these sit alongside.
Two cities bought the same acoustic gunshot-detection system and inherited the same criticisms. Only one of them commissioned the study that could settle the argument — and that, not the hardware, decided what happened next.
| Project | Since | Status | Outcome | Evidence | Run by | Cost | What settled it |
|---|---|---|---|---|---|---|---|
| ShotSpotter Acoustic Gunshot DetectionChicago, 🇺🇸 | 2012 | Completed | Failed | Independent | City government, Private company | ≈$53M total contract value (initially $33M for 2018-2021) | The city's own Inspector General reviewed it, the contract was allowed to lapse, and police stopped using the system in September 2024. |
| ShotSpotter Acoustic Gunshot Detection, In Service Since 2006Washington D.C., 🇺🇸 | 2006 | Live | Ongoing | Media | City government, Private company | At least $5.16M paid to SoundThinking since 2016; no current contract value or term is public | The same sensors, running since 2006 and still running: no equivalent review was ever commissioned, so nothing forced a decision. |