I can’t help but think of the street person with a sign saying, “Will work for food” when I read about an experiment conducted at bus stops in Finland.
Vassilis Kostakos at the University of Oulu in Finland and colleagues put up four large LCD touchscreens in bus stops and other busy public areas. A sign invited passersby to touch — and to help train malaria-detecting software.
Was it worthwhile? Apparently it was.
They got 1200 answers over 25 days and found the accuracy of the results was comparable to those of paid workers on Mechanical Turk. Lone users were the most active, perhaps because they were free from peer pressure or distraction. These loners also created a honeypot effect, attracting more users, the researchers will tell the UbiComp conference in Zurich, Switzerland, in September.
Vassilis says such displays could [also] be used to tap local knowledge, such as the best place to walk a dog, the meaning of some confusing signage, or what bands are playing in town.
The researchers promised nothing to participants, of course. But just wait for this to catch the eye of alert transit administrators, and for transit districts to adopt it.
Where there’s machine learning there’s money. Starving public transit should have some of it, especially if it works for food.
Leave a Reply