Data Smellification — Smelling Data Saved My Life

• Read original article here

We usually associate our sense of vision to how we interpret data, but we can consume data with other senses, too

A few weeks after moving into a new, old house and working from home all day, I felt sick. I called my husband who was traveling, and begged him to come home early for the weekend.

He walked in the door at 11pm, and announced “it smells like gas in here.” We called the gas company, and the technician found 4 gas leaks throughout the house. I couldn’t detect the small increase in the amount of natural gas, but coming in from fresh air, my husband could. He probably saved my life.

Natural gas is one of a few instances of man-made data smellification. Mercaptan liquid is added into the naturally odorless natural gas so that leaks can be readily detected. It was because of this smellification process that my husband could notice the leak. Without this intentional addition of scent, natural gas leaks would be impossible for the average consumer to detect.

There are many natural data smells that we use to avoid harm. The smell of food provides us with data about its appeal, nutrition, or toxicity. For example, the smell of burning food tells us how our cooking is going (badly), while the smell of rotting food tells us it is time to clean out the fridge. Indeed, the very fact that certain things smell good or revolting to us is the result of thousands of years of evolutionary selection: things which are dangerous to us tend to smell bad, and things that are good for us tend to smell good.

This is not the only common case of data smellification, however. In fact, we often add smell to things to give “false data” or to give the impression that something is better than it actually is.

For example, before going in for a kiss, you might pop a stick of mint gum or a swig of mouthwash to freshen your breath. The scent signals real data: hygiene level. You want to appear clean, like you just brushed your teeth, even though that may not be true. There’s a similar motivation for applying perfume or cologne before going out on a date. These things cover up the true data (your scent), which might be less appealing.

Air fresheners are another example of this phenomenon, specifically ones that convey freshness. Here, natural smell provides the true data, “how clean is this car, or how new is this car.” Air fresheners attempt to falsify that data. The car is not going to smell like a “new car” no matter how many little “new car smell” trees you hang from your rear view mirror.

Leanne Wijnsma and Froukje Tan started The Smell of Data project to alert internet users of data leaks on personal devices using smell. I’m excited by this project of adding smell as a way to detect the presence of information. We don’t use our sense of smell as often as our other senses, and this project uniquely harnesses that with a custom scent.

We spend a lot of time thinking about visualizing data here at the Data Visualization Society, and rightly so; it is a great way to use our one of our most powerful senses (sight!) to make sense of data! However, I’m intrigued by moving past just the visualization of data and into other senses, like sound (TwoTone: a data sonification app and Loud Numbers: a data sonification podcast), touch (dataphys.org: a list of physical visualizations and DayDohViz: play-doh visualizations), and particularly smell!