Ready for a smellfie? Photo tags map the smellscape of a ci

General Discussions
Post Reply
User avatar
Nipuna
Moderator
Moderator
Posts: 2729
Joined: Mon Jan 04, 2010 8:02 pm
Location: Deraniyagala,SRI LANKA

Ready for a smellfie? Photo tags map the smellscape of a ci

Post by Nipuna » Sat Jun 06, 2015 11:55 am

You could use the maps find the sweetest-smelling route around London (Image: Eye Ubiquitous/REX Shutterstock)
You could use the maps find the sweetest-smelling route around London (Image: Eye Ubiquitous/REX Shutterstock)
dn27629-1_300[1].jpg (18.38 KiB) Viewed 17503 times
Are you ready for your smellfie? Photo captions on visual social networks like Flickr and Instagram can be used to create detailed smell maps of urban areas.

A team led by Daniele Quercia at the University of Cambridge used data taken from the sites to plot maps of aroma distributions across whole towns and cities, from the stench of sewer vents and horse dung to fragrant fruit markets, the new-mown lawns of public parks and the garlicky whiff of restaurant districts.

The system might lead to some unusual location-based technology: automatically generated online maps that let us plot routes not by shortest distance, but by the smells we prefer to experience (or avoid) on a walk, jog or bike ride. The name for these personalised maps? "Smellfies".

The research is based on the manually produced urban smell maps and aroma lexicon developed over the last three years by artist and designer Kate McLean of the Royal College of Art in London - a member of Quercia's team.

McLean has been taking volunteers around cities like New York and Amsterdam, getting them to agree on smells at certain points - sniffing everything from garbage to manhole covers to park benches - and plotting maps. This enabled her to create a "smell dictionary" of 285 words commonly used to describe urban smells. But mapping them manually is slow work.

The researchers reasoned that the captions and tags that people add to their geotagged online photos on Twitter, Instagram and Flickr might offer a faster way to build smell maps.
(Image: Daniele Quercia. Key: Orange/red: Low/high smell of vehicle emissions; light green/dark green: low/high smell of flowers, plants and trees)
(Image: Daniele Quercia. Key: Orange/red: Low/high smell of vehicle emissions; light green/dark green: low/high smell of flowers, plants and trees)
dn27629-2_1200[1].jpg (475.22 KiB) Viewed 17503 times
Map showing smells of vehicle emissions and plants, flowers and trees in Barcelona

So for Barcelona (map above) and London (see below), the team collected geo-referenced tags from about half a million Flickr pictures and 35,000 Instagram photos as well as 113,000 tweets. The tags and tweets were matched with words in the smell dictionary.

"We simply looked for the keywords Kate has identified in the picture captions and tags on their photos," Quercia says. The smell maps were presented at a conference in Oxford this week.

Flickr and Instagram worked best, probably because they have a formal captioning and tagging structure, whereas tweets mentioning smells tended to be far more general in scope and not as area-specific.
(Image: Daniele Quercia)
(Image: Daniele Quercia)
dn27629-3_1200[1].jpg (563.74 KiB) Viewed 17503 times
London's animal smells, shown in blue (eg London Zoo, horses and dogs in Hyde Park)

The maps provide an entire "smellscape" of a city, and one that can be changed to plot the smells most apparent at different times of the year, which might be useful for tourism planning.

"We also tried correlating the data with known air quality indicator data in London and found that pleasant nature-related smells were suppressed by air pollution," says Quercia. The team hope the findings will help urban planners better manage air quality and expand their monitoring of airborne pollutants.

The team plan to launch Smelly Maps, a website where you can tailor your own olfactory experience, or smellfie, in the next six months.
Post Reply

Return to “General Discussions”