Discovering (or Disproving) Aliens with Data Science

Use Cases & Projects Rebecka Flynn

Many real Americans believe they have seen aliens, and we have the reporting data of UFO sightings to prove it. The dataset records the time, date, location, shape of the sighting, and the story the reporter submitted of their sighting from the early 1900s to 2014. So I dove into the data to see what other interesting insights we could find.


I don’t know whether aliens are real or not, or if they actually visit cornfields. I’m definitely more of a Scully than a Mulder, and I did get excited by the sighting—I mean explosion, in New York in December. But let the record show: I have no coding skills. I’ve tried python and R before, but I don’t call myself a data scientist. As a result, I used a point-and-click interface for all of my analysis in this project without the addition of code.

I combined the alien sighting data with 2010 census datasets on county poverty and population rates as well as US military installation locations. Since the majority of sightings occurred after 2000, this dataset combination seemed apt.

military installations and alien sightings

aliens v. military basesWhile I did not find a correlation between poverty levels and UFO sightings, I did notice a strong association between military bases and UFO sightings. The map above shows the top counties for sightings, color-coded by their distance to the closest military installation. Most are within 15 miles of a military installation, with few outliers. We found that 61 percent of all UFO sightings are within 24 miles of a military installation, and 82 percent are within 42 miles.

This chart below shows the type of sighting (shape) compared to the frequency (color-coded) and distance from military installations (join_distance). As you can see, the majority of the finding are within 30 miles from an installation. The predominance of “circles,” “fireballs,” “lights,” and “Others,” (meaning other-worldly) don’t seem inconsistent to me with helicopters dropping flares or doing maneuvers.

aliens by shape

Where do aliens hang out?

There were some spots where alien sightings were reported at much higher rates than the population should warrant. After removing outlier counties with populations below 10,000, four of the top 10 counties are in Washington state. I have no concrete explanation for this trend. Perhaps the state’s low-light pollution and relatively high latitude mean that residents see the northern lights as UFOs, or maybe aliens are like Twilight vampires and just enjoy the rain.

The Truth is Out There

Holiday festivities or aliens: you decide. When we look at popular dates for reporting UFO sightings, a hilarious trend emerges. The most common day to see a UFO in America is July 4th, and the second most popular time is at New Years.month and date frequency

Considering the popular shape keywords (light, circle, fireball) from the chart above, and Americans’ propensity to set off fireworks on those dates, it seems to me like fireworks are being mistaken for UFOs.

I will never say aliens do or do not exist. But I will say, it is very funny that the day we are most likely to see aliens is also the day we are most likely to set off fireworks.

death star
If you want to try your own code-free analysis of strange phenomena, check out a free Kaggle dataset and get a free Dataiku license to perform fast robust analysis.

Try Dataiku

You May Also Like

Top High-Value AI Use Cases in Retail & CPG

Read More

How AI Is Transforming the Wealth Management Industry

Read More

Two Large Barriers to Widescale AI Adoption in Chemical Manufacturing

Read More

Military IT: The Three Best Ways to Scale & Mature

Read More