I built a public-source evidence surface to test one narrow UAP claim:
Do public UFO/UAP report rows cluster near nuclear power plants?
This is not a truth claim about any specific sighting. It is only a geography/proximity test using public report rows, Census place population, major airport proximity, and matched non-nuclear power plant controls.
The result surprised me a bit:
The nuclear-specific proximity claim did not hold in this dataset.
Primary 50-mile test:
- Public geocoded report rows: 105,250
- Nuclear power-plant sites: 57
- Matched non-nuclear power-plant controls: 285
- Mean reports within 50 miles of nuclear sites: 649.98
- Mean reports within 50 miles of matched non-nuclear controls: 763.05
- Nuclear/control ratio: 0.851823
- One-sided p-value for nuclear > controls: 0.854146
So in this reduced public dataset, nuclear sites did not show a stronger report-density signal than matched non-nuclear power sites.
The stronger visible pattern was reporting geography.
Population mattered a lot:
- 105,086 of 105,250 geocoded rows matched to Census population places
- log population vs. log report-count correlation: 0.641288
Airport proximity was also measurable, but it mostly followed population geography:
- Median report-row distance to a major scheduled airport: 13.277 miles
- Population-weighted Census-place median: 10.764 miles
So airport proximity looks like a confounder, not a clean standalone explanation.
Important limits:
- This does not prove any sighting is true or false.
- This does not prove aircraft explain every report.
- This does not test classified military/nuclear weapons sites.
- This uses Census place centroids, not exact witness GPS.
- 17,783 U.S. rows did not resolve cleanly to Census place centroids and were excluded from spatial tests until recovered.
What I think this shows:
Public UAP report geography is heavily shaped by where people live, report, and have open sky visibility. If someone wants to argue for a nuclear-specific signal, the comparison needs matched controls, not just “there are reports near nuclear sites.”
This was built as an evidence surface, not a conclusion machine. The useful part is that every claim stays bounded: what the public data supports, what it does not support, and what needs another run.
HF space link: please let me know if you see anything wrong, or if I can add in any tests, using public data only.
https://huggingface.co/spaces/cjc0013/nuclear-uap-evidence-surface
by Either_Pound1986
3 Comments
Good job. THIS is science.
So basically you found that the higher the population density the more likely it is that you get UAP reports. Which is to be expected. And also found there is no correlation with nuclear sites contrary to a narrative which is out there.
Excellent job! Data analytics are the key to separating the signal from the noise.
I’m honestly a little surprised. I assumed the underlying premise (more UFO reports at nuclear sites) was true, but my working hypothesis was that it was not because UFOs appeared more often but because of better security looking outward and upward.
Now it seems as if that premise was flawed: UFOs don’t appear more often.
Time to reconsider my priors!
I think the default mindset is surprise , “I just saw .. you won’t believe it” Sightings in the wild are hide in plain sight and around sensitive areas meant to say “we see you”