The 10 Best Data Science Halloween Costumes

Posted 2018-10-15T Posted by Tom
Wow your friends and colleagues with these 10 creative data science inspired Halloween outfits.

Random Forest

This one only works if you have some data science companions who can join you in the ensemble. For this costume to be true to its name, you’re going to need to ensure and verify its randomness. I recommend waking up a little earlier on the day and enshrouding yourself in hoards of double-sided sticky tape. Then, take advantage of the falling autumn leaves by diving into some impromptu forward rolls on your daily commute, thus, adorning you with an indiscriminate orange and brown collage, and potentially the odd chocolate bar wrapper, which would allow for a powerful statement about the environment. Either way, you must agree this is pretty random, right?


Dress up as a famous Julia, such as Julia Roberts, or…… So, dress up as Julia Roberts for the day or if your name already is Julia, just show up to work as usual and remind everyone of their ignorance when they suggest you “haven’t bothered to make an effort.” Tut, tut, tut.

Supervised Machine Learning

Grab a few arbitrarily cut strips of tin foil and get them stuck on your shoulder blade, knee-cap, and forehead for the mainstream cyborg look. Make sure to enlist the help of a colleague for this one and instruct them that they will need to label all the equipment you’ll be using, the food you’ll be eating, and so on.

Unsupervised Machine Learning

Ditch the colleague.

Deep Learning

*See the two examples above* I don’t advise crouching down for the entirety of the day, so be sure to bring a shovel, snorkel, and play sullen Drake hits throughout the day to illustrate just how deep your learning is.


Halloween falls on a Wednesday this year. Show the pitfalls of extrapolation by wearing your Monday, Tuesday, Wednesday etc… socks on their respective days the week before and then boom, wear an odd pair of the most eccentric socks in the observable universe on All Hallows Eve.

Decision Tree

This one requires a lot of (paper) straws. Find an old jumper and cut and stick one straw leading to two straws, leading to another two, and so on. Alternatively, you could utilise a few sticky notes and pen a few questions regarding workplace matters, such as, “Is it an emergency?”, with the answer, “No”, leading to the statement, “Then don’t disturb me.” If any of your colleagues start to question this seemingly passive-aggressive attire, just remind them that it’s all for show, part of the costume, and any resemblance this bears to things you were moaning about last week is simply a coincidence. Correlation doesn’t imply causation, remember.


Draw a Wi-Fi symbol on everything, including your forehead. Episodically throughout the day, freeze mid-way through a task or conversation, and maybe give buffering a go?

Data Lake

Adorn yourself head-to-toe in sparkling blue; the kind of blue you see in movies but seldom in real-life. Remember to slap on some ones and zeros to ensure everyone knows you aren’t just any old lake. Depending on how laid back your office is, you could always be very direct to the point by bringing in a miniature water gun. Too much?


So, it’s the obvious one, but simply dressing up like a snake isn’t going to cut it. That’s more reptile than data science, so you’re going to have to behave like Python as well as look like it. This means yelling, “Invalid Syntax,” anytime you disagree with a colleague, refusing to run any errands if someone is less or more than four spaces away from you, and suffering from a few runtime errors, resulting in you having to nip outside the office for another lunch break whilst your colleagues attempt to determine why.

Get priority access to Pivigo news, features, events and networking opportunities