Close Menu
Finding Patterns in 152,000 UFO/UAP Sightings
UFO

Finding Patterns in 152,000 UFO/UAP Sightings


I used ML/AI to analyze the NUFORC sightings database. Some recurring patterns:

  • Triangles almost never appear during the day
  • Disks are 5x more likely to involve entity encounters
  • The majority of sightings are silent

Full analysis and methodology here.

by deckerRTM

View 4 Comments

4 Comments

  1. deckerRTM on

    I work in data professionally and have been applying ML clustering techniques to the NUFORC database as a side project. The article goes deeper into the methodology and breaks down patterns by shape, time of day, and encounter type. Interested in any feedback

  2. The_Great_Dadsby on

    I absolutely love this!!

    I have my own UAP “AI” project that I’m working on. It’s more focused on the “woo” side of things and includes all sorts of books, papers, studies, interviews etc from religious texts (not just the abrahamic religions), to folk tales to Remote viewing, old occult literature, astral projection etc. I’m definitely going to try and learn some lessons from your hard work!

  3. Antique-Wonk on

    Doing some visualisation and correlation stuff at the moment. TBH it’s the work of Clawdbot. My bro and I just told it what we wanted and it served up and interactive web page. Still going through it to check nothing silly has happened. We put a time slider on it and an events log (wars, nukes, catastrophe, etc) so we could see if there were any correlations. One correlation I think we’ve found is on M5 plus earthquakes and UFO sightings around them. 90% of the time. Still checking the data so it might be bogus yet.

  4. shancats on

    First of all I want to say great work! I’ve been thinking about using AI to classify the body of UFO reports. I hope to see more of this kind of work from people.

    It would be interesting to use AI to classify and add metadata to reports for further meta-analysis. I’m thinking things like adding various tags to reports e.g., experiencer sentiment (bliss, peace, love, experimentation, evil, malevolence, etc.). Coming up with categories and tags seems like a task in itself but once done it would be interesting to build some kind of visual that could hopefully make the data clearer.

    Nitpicks:
    The tables throughout (The Sound of Silence, Disks are Different, etc) are black text on a dark background and is quite difficult to read.