Half Baked Fixes
Published in

Half Baked Fixes

Driver Rudeness vs. Vehicle Selection

Are Audi drivers actually the rudest on the road? Let’s measure it!

Whenever I get cut off by a BMW or Porsche (or maybe a Model 3), the first thought that pops into my biased brain is: “Well that figures.”

Again, it could just be bias. But then again, what if it isn’t? Could one design a system to measure “motorist empathy” and answer once and for all whether rude people are attracted to certain brands?

One could. One shall.

First: who cares?

Brands that are associated with friendly drivers, for one. Wouldn’t it be nice if, I don’t know, Chevy Bolt drivers knew for a fact that we, whoops: they, are the friendliest drivers on the road? That might garner some loyalty, or attract other friendly drivers.

OK, on to business.

The Solution: a Driver Rudeness Score based on driving habits associated with rudeness

First, we’d need to establish some metrics which are objective and consistently measured.

Here are two:

- rapid lane changes per mile driven (i.e., “cut off rate”)

- average following distance in meters (i.e., tailgating)

Both of those could be measured using the accelerometer in any modern phone. You could build an app to do this, which just runs in the background.

Let’s define each metric.

  1. A rapid lane change is a sudden lateral acceleration followed by rapid correction in the other direction. I’m sure you could build some rules around this, or you could use deep learning like any modern engineer would.
  2. Average following distance would ideally tap into the vehicle’s radar or vision system (if it has one), which is entirely impractical. A reasonable analog is braking behavior: the more often you brake, the closer you are to the car in front. I would measure this based on braking frequency above a given speed. If you brake constantly without coming to a stop, you are probably tailgating.

Both metrics would require some calibrating, but it’s a start.

Once the metrics are established, you could measure them for, say, 100 drivers for each make and model of interest for 6 months.

The worst make and model would be determined by taking the median value of its 100 drivers for each metric, then normalizing the median values across the other makes and models and ranking them. The brand with the lowest average ranking between the two metrics would by the friendliest. The highest average would be the rudest.

Here’s to hope, Audi drivers.




Half Baked Fixes identifies everyday issues (market failures, operational inefficiencies) and provides a framework for solutions, some of which may surprise you.

Recommended from Medium

How to become a data scientist?

The Statistics of Persuasion

Chart It, Don’t Chat It: Data Visualization Tips.

Power BI Job Roles and Responsibilities

Interview #12: Orthi Rabbane, Data Science & Data Strategy

Implementation Linear Regression — the price of coffee and the diameter of the glass | python.

Building tools to bring data-driven reporting to more newsrooms

Part 1 : Introduction to Time series Analysis

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Pat Canny

Pat Canny

More from Medium

Embryo selection for IQ and other ‘marginal’ benefits

“War is in the Air”

Did you know? … eight entertaining facts about board games!

The 4 Most Common Geometries in KML