A SILLY ML PROJECT

We made ML models for the dumb things you do every day.

Study or scroll? Quick nap or sleeping till dinner? Fix the bug or restart the server? There's a model for each.

Try the experiments →How it works

INPUT

deadline: tomorrow
sleep: 5.2h
phone: in hand
last activity: scrolling

MODEL

HUMAN-001 · ...

PREDICTION

running…

WHY?

Machine learning solves serious problems all day.
We gave it a break.

The experiments

01Productivity
Will I study or open Instagram?
A predictive model for the timeless battle between productivity and scrolling.
8 inputs · Logistic Regression
Run experiment
02Food
Will this Maggi become too spicy?
Machine learning finally answers the question nobody asked.
5 inputs · Random Forest Regressor
Run experiment
03Sleep
If I lie down for 5 minutes, will I wake up?
An advanced system for predicting the consequences of "I'll just rest for five minutes."
8 inputs · Logistic Regression
Run experiment
04Digital habits
Which app will I open next?
Predicting the next digital destination of a chronically distracted human.
10 inputs · Logistic Regression
Run experiment
05Daily life
Will I take my phone to the bathroom?
A classifier designed to answer one of humanity's most avoidable questions.
6 inputs · Logistic Regression
Run experiment
06Engineering
Will this programmer fix the bug or restart the server?
Predicting the oldest debugging strategy known to mankind.
7 inputs · Logistic Regression
Run experiment
07Relationships
Will she reply or am I cooked?
An intentionally experimental model attempting to predict reply behavior.
7 inputs · Logistic Regression
Run experiment
08Academia
What excuse will I make tomorrow?
Machine learning meets academic procrastination.
6 inputs · Logistic Regression
Run experiment

How it works

The data is made up and the models learn from it like any other data. Nothing is trained on this page, it just runs the finished model.

01
Human behavior
Something small, silly and very relatable.
02
Data
Made-up data with a bit of noise.
03
Features
Raw inputs like time and apps turned into numbers.
04
Model
Logistic regression or random forest, whichever did better.
05
Prediction
A probability, right on the page.

See how it's made

MODELS: 8
DATASETS: 8
ROWS OF DATA: 85,000

About the experiment

I built this to turn silly everyday habits into machine learning problems.

Every model has its own notebook and dataset. The site just shows what comes out of them.

More about the project