Jetpac's Peter Warden analysed 100m photos from Instagram's public system, as part of the company's attempt to build a recommendation system built purely on pictures which are geotagged.
The happiest place in the UK? It's easy to find: all you have to do is analyse the colours, facial expressions and other objects in tens of millions of location-tagged photos posted on Instagram. And it turns out that the happiest city is Belfast - and the happiest place there is a pub called the Parlour Bar in Elmwood Avenue.
Why? Because the people in photos posted from around that location tend to be smiling, and few look grumpy.
The least happy place, meanwhile, turns out to be Salford, which comes below London and Bath in an analysis of 40 cities by Peter Warden, co-founder of the UK startup Jetpac, which provides guides of places to visit around the world based on publicly posted pictures.
Warden analysed 100m photos from Instagram's public system, as part of the company's attempt to build a recommendation system built purely on pictures which are 'geotagged' - linked to a specific location. He got software to analyse the faces in the pictures, which first found the mouths of people in the pictures, and then decided - based on colour - whether they were wearing lipstick (which would indicate being 'glammed up', and so likely to be having a good time) and whether they were smiling - which is rated on a scale of 0 to 100.
The software could also identify moustaches - and so could point to the places in cities likely to have the largest concentration of 'hipsters' - although, Warden notes, they tend not to smile as much as others. (That could be part of the reason why London rates so lowly on the top 40.)
Using that system, Belfast turns out to be the happiest city - and the Parlour Bar, close by the Methodist College and the Botanical Gardens, is ranked as the ground zero of happy people.
'We get a lot of students,' said John Kennedy, the manager at the bar, when contaced by the Guardian. 'They bring a fun element to the bar. And we've got a quirky atmosphere. Plus oven-baked pizza.' The 550-capacity location also means that students armed with smartphones are highly likely to take and share photos from there.
Warden's analysis was made possible because we are generating so many photographs every day. More than a billion photos are taken worldwide on a daily basis. About 55m of those are taken on instagram - and of those, 30% are public, and of those 30% are geotagged, yielding nearly 5m photos every day on Instagram alone which are public and geotagged.
'We're identifying particular venues (bars, hotels, parks, etc) that we are interested in, and then we query Instagram for public photos at that place over the last year,' explains Warden. 'Popular places will have hundreds or thousands of photos, and we've looked at over 6 million venues so far, so the numbers add up pretty quickly.'
Even with a few hundred photos from a location he says it's possible to produce 'reasonable data' on gender ratios, popular hours, the moustache density, coffee cup prevalence (indicating whether a place is a coffee bar or bar) and so on.
'It's also possible to look at where else the photographers at a place take photos, so we can tell if a bar's popular with students by the number of photos at university buildings the people there have taken,' Warden explains.
Jetpac's systems have analysed photos to create many other categories - including intellectuals (art galleries and museums), jet-setters (take many airport building photos), parents (photos from nurseries or toy stores), musos (photos from concert halls and music festivals), sports fans and more.
Warden, who formerly worked at Apple, said: 'I've been searching for an interesting and accessible data source like this for years, so it's been great to find one I can turn into a useful product. I'm convinced all the 'data exhaust' people are generating can actually help us in a lot of ways, so it's great to get beyond Tomorrow's World tech demos and have something a bit more down-to-earth to show.'
UK cities' happiness ranking
1 Belfast 34.292 Bournemouth 32.883 Derby 32.604 Exeter 32.345 Lincoln 31.956 Aberdeen 31.147 Glasgow 31.028 Colchester 30.689 Newcastle upon Tyne 30.4310 Oxford 29.9611 Coventry 29.8912 Cardiff 29.7913 Wolverhampton 29.6514 Milton Keynes 28.9715 Canterbury 28.9716 Leeds 28.6217 Liverpool 27.3918 Cambridge 27.3019 Manchester 27.2020 Plymouth 27.1821 Birmingham 26.9522 Leicester 26.2923 Portsmouth 26.2024 Windsor 26.1425 Reading 25.9426 Southampton 25.8627 Newport 25.0928 Nottingham 24.9129 Bristol 24.2130 Norwich 24.0731 Blackpool 23.9632 York 23.4433 Brighton 23.4434 Sheffield 23.3835 Newquay 22.9636 Edinburgh 22.3137 Luton 22.2238 London 20.4839 Bath 18.3240 Salford 17.90
Tidak ada komentar :
Posting Komentar