Disentangling EV Charging Behavior: An Integrated and Explainable Framework for Predicting Session Duration and Energy
Every evening, across cities, thousands of people plug in their electric vehicles. Some do it as soon as they reach home. Some wait until late night. Some forget and charge only when the battery is almost empty. At first, this may look like a simple routine.But if you look closely, every charging action tells a story. And that story is not just about machines—it is about people.
A Simple Question That Isn’t So Simple
We started with a basic question:
Can we predict how people charge their electric vehicles?
It sounds easy. But the moment we looked at real data, things became complicated. Two people with the same car behaved completely differently. One charged daily. Another charged once in three days. One unplugged quickly. Another left the car connected for hours.
That’s when we realized:
EV charging is not just a technical problem. It is a human behavior problem.
Listening to the Data
Instead of forcing all users into one pattern, we decided to listen to the data. We used Artificial Intelligence to observe how people actually behave. Slowly, patterns started to appear. We could see different types of users emerging naturally—almost like personalities.
Meet the EV Users
From the data, four types of users stood out:
- The one who charges every day after work
- The one who prefers charging overnight
- The one who drives long distances and charges less often
- The one who plugs in and forgets, leaving the car connected for long hours
These are not just categories. They are real habits. Real lifestyles. And suddenly, EV charging started to feel more human.
A Surprising Realization
As we went deeper, something interesting happened.
We were trying to predict two things:
- How long the car stays plugged in
- How much energy it consumes
We expected both to behave similarly.
But they didn’t.
1. Charging time was about people
It depended on habits, routines, and time of day.
2. Energy used was about the machine
It depended on battery size and vehicle design. This was a turning point. It showed us that:
Some problems need human understanding, and some need engineering logic.
Why This Matters in Real Life
This understanding can change how we design systems.
- Power companies can better manage electricity demand
- Charging stations can reduce waiting time
- Users can get smarter charging suggestions
What This Journey Taught Us
At the beginning, we were trying to build a prediction model. But in the end, we learned something deeper:
Data is not just numbers. It is a reflection of human behavior.
And when we combine AI with that understanding, we don’t just predict actions—we understand them. Every time someone plugs in their EV, it may look like a small action. But behind it lies a decision, a habit, a pattern.And when we start understanding these patterns, we move one step closer to building a smarter, more human-centered future.