The UAP Atlas
156,000+ UAP / UFO reports mapped across a century — with the ordinary explanations checked, not hyped. Reports, not confirmed sightings.
We rate the hype — we don’t spread it.
Bright planets, the moon, meteor showers and rocket launches — next to the month’s report count. New every month.
Where people look, and what they see
A cross-section of every report in the atlas, pooled from NUFORC and a global report catalogue. Four questions: which places report the most, which shapes keep coming back, how those shapes shifted, and when the sightings happen.
Reporting hotspots
The source data has no place or region field, so each report with coordinates is assigned to the nearest major metro (within 120 km). What remains is an unmistakable North-American centre of gravity — with London, Toronto and Vancouver the outliers beyond it.
- New York16,971
- Los Angeles16,589
- Boston9,716
- Seattle8,313
- Chicago7,915
- San Francisco7,553
- London6,626
- Phoenix6,370
- Columbus OH6,309
- Denver6,285
- Toronto6,268
- Detroit5,784
- Portland OR5,515
- Washington DC5,144
- Baltimore4,952
Reported shapes
The 30 raw shape labels are grouped into families. About 40% of all reports carry no shape, so the split below covers the 269,889 reports with a recognisable shape. The images are AI impressions of each family — illustrative, not real captures. Tap any to enlarge.
The beings people describe
A small fraction of reports describe not just an object but its occupants — the Hynek scheme codes 14,260 of them as close encounters of the third kind. Across decades a handful of descriptions keep recurring. This is a guide to those recurring types — a taxonomy of what witnesses describe and how the culture pictures it, not a catalogue of real creatures.
These are recurring shapes in testimony and folklore — how people describe and picture occupants, not a verified list of real beings.
Each image is an illustration of the described type, generated to a neutral brief. None is a photograph or a real capture.
The types track media as much as sightings: the Grey took over only after the 1960s, and later additions like the Reptilian arrive with their own era.
When they appear
Timing is the strongest signal in the whole set: sightings cluster after dark and swell in summer — exactly when people are outside looking up.
Reports pile up between 19:00 and midnight. The 0:00 and 12:00 bars are inflated by reports logged with a placeholder time.
The usual suspects
Before it’s anything exotic — is it the sky itself? Three checks against the objects people most often mistake for a craft.
Each phase is an equal slice of the ~29-day cycle, so an even sky gives 12.5% apiece. Bars show the deviation from that even split.
- New12.58%
- Waxing crescent12.54%
- First quarter12.70%
- Waxing gibbous11.86%
- Full12.21%
- Waning gibbous12.71%
- Last quarter12.85%
- Waning crescent12.55%
Reports barely track the moon: every phase sits within ~0.6 point of even and mean illumination is 0.50 — no dark- or bright-sky pull. There is a faint dip around the bright waxing-gibbous/full moon, but with 450k dates even that tiny swing is “statistically significant” — a reminder that effect size, not just p, is what matters.
Venus is the sky’s brightest point and the textbook “UFO”. Bars show the share of reports during each apparition; the tick marks the share of calendar days (chance). Reports run above chance when Venus is an evening star, below when it is a morning star.
- Evening starup after sunset36.3%z+23
- Morning starup before dawn33.1%z-18
- Near the Sunlost in the glare30.7%z-5
A small but consistent lean (~5% relative), matching the evening report peak — Venus is the classic misidentification. Still only a correlation: it doesn’t tell us which individual reports were the planet.
The second-brightest planet, up much of the night around opposition. The same test as Venus — reports vs calendar days for each apparition.
- Evening skyup after sunset42.2%z-15
- Morning skyup before dawn44.9%z+19
- Near the Sunlost in the glare12.9%z-6
Jupiter’s lean is faint and, if anything, tips toward the morning — no clean evening signal. Higher and far less brilliant than Venus, it makes a much less obvious “UFO”.
How the shapes shifted
The same families, plotted by decade. The flying saucer ruled the 1950s and 60s; since the 1990s, orbs, lights and triangles take over — the age of the phone camera and the drone shows up in the reports.
More in the data
A few extra cuts worth a look — how cases are classified, which day and country they come from, the biggest years, and whether the reported shape changes around the world.
The classic scheme: distant lights and discs, then “close encounters” graded by proximity and interaction. Nocturnal lights dominate.
- NLNocturnal light38.2%
- CE1Close encounter I · near sighting13.4%
- DDDaylight disc12.0%
- CE3Close encounter III · occupants10.2%
- CE2Close encounter II · physical trace10.0%
- CE4Close encounter IV · abduction2.8%
- RVRadar-visual1.8%
- ·Other / specialised codes11.6%
- Canada5,722
- United Kingdom3,377
- Australia918
- India492
- Mexico404
- Germany220
- New Zealand170
- South Africa168
- Netherlands164
- Ireland160
Among records that carry a country, the USA is ~90%. Most non-US catalogue rows have no country tag at all, so this under-counts the rest of the world.
- 201414,517
- 196714,489
- 201214,313
- 201313,341
- 201512,940
- 202012,754
1967 was a classic wave; the 2012–2015 surge is the smartphone era — more cameras, more reports.
- USALight 22%Orb 21%Triangle 11%
- UKOrb 23%Light 21%Triangle 10%
- CanadaLight 24%Orb 20%Triangle 11%
Light, orb and triangle lead almost everywhere — the reported shape barely shifts by country.
Hotspots track where people live and report. Dense, English-speaking areas are over-represented — this is a map of attention, not of arrivals.
NUFORC is heavily US-focused; the global catalogue fills in the rest of the world but often leaves country and shape blank. Hence the large “unknown” share.
Cities are matched by proximity to ~75 world metros; reports more than 120 km from one are left out of the hotspot chart.
These are reports, not confirmed sightings. Report counts track population and media attention more than any phenomenon, and coverage is uneven (NUFORC is US-heavy). How to read this →
156,000+ reports on a world map — filter by year and shape, overlay quakes, launches and lightning, test correlations.
Landmark casesThe most-documented events — hearings, military encounters, mass sightings — each with its primary source.
MethodologyReporting bias, population normalization and the null model — the rules that keep this honest.
Open dataDownload the reports as CSV or JSON, or call the open API — factual fields with a link back to each source.
Report a sightingThree steps, with a live sky check for launches, planets and weather before you submit.