Using data from geo-tagging to map the Happy City

photo_sharing_logos

by Irina Shafranskaya*

Photo-sharing is currently becoming a huge part of social media activity. Several applications, with Instagram the most popular among them, represent people’s emotions. Such data pose new challenges for city data analysts as a lot of pictures are geo-tagged. City representation via images is not a new topic; it seems to us that Antonioni was one of the first with his “Blow-up”, who tried to catch the place by a camera click in his 1966 film Blowup. The digital era just brings new insights – as Ames and Naaman (2007) argued. Instagram covers additional aspects of this representation as sociality and functionality – we geo-tag places to give a special social signal of the places’ livability and share our emotional state-of-the-moment.
The concept of participatory sensing developed by CENS (UCLA) could help to understand our cities better in order to improve them, and our project was aimed to evaluate city happiness using geo-tagged Instagram images. City happiness is in fashion nowadays as the alternative approach to city comparison: Happiness Index is usually calculated by a combination of statistics, survey results etc. Yet, we doubt that numbers properly reflect the idea of happiness, which is quite situational and spontaneous. Contrary to measurements we try to define city happiness using crowdsourced geo-tagged city images. First, we downloaded all the geo-tagged images of Perm (Russia), the city we live in (more than 27 thousand images) and put them on the city map – it appears that the whole city is totally covered by users’ sharing activity.

Irina1

Then we developed a metric for happiness evaluation using two tools: the evaluation of the content of the images’ comments (from negative to positive) and the evaluation of peoples’ emotions on the pictures (using Microsoft Emotion Recognition service). This finally gave us the metric for every picture – from 0 to 1 – the happier both picture and comment are, the closer it is indexed to 1. The picture below illustrates the emotional map of our city Perm. The closer the metric to 1, the darker the colour blue.

Irina2The city seems to be quite happy, but several locations have a lower estimation in our “happiness index”, which is based totally on emotional representation. Locations with low numbers are also visible – isn’t it the better metric for the quality of the geo-tagged place than any statistical one? We focused our attention on several things, for instance, on the prominent tourist walks, officially routed by the city authorities.

Irina3Surprisingly, we have found out that these walks are not popular as places to be pictured and tagged – isn’t it again an issue for those who create these routes without taking into account a participatory approach? How exactly are they organized? And are they really the points of interest? Before we had such data, we were sure they were, but now we have good reason to doubt. Similarly, we combined the main city flagship tourist objects (orange dots on the picture below) with Instagram images, and again they appeared to be not as interesting in terms of photo-sharing and not quite as happy as some other places of the city.

irina4

Crowds have their own wisdom, and though our research we show that this should be taken into account. The city can have its emotional scape, which changes from moment to moment, as the amount of data grows exponentially. City emotions could hardly be grasped at every moment, but every digital snapshot contributes to the whole sense of the city, and every social media user participates in city representation, geo-tagging her or his picture.

So smile, you could make your city happier!


*Irina Shafranskaya

Irina Shafranskaya, is Associate Professor at the Higher School of Economics, Russia.Her primary research interest is how city branding concepts are applied in a challenging Russian context. What inspires her most is the understanding that city branding is a powerful tool for changing residents’ perceptions – which allows us to making our cities better.

The Instagram data was processed by Mark Evlampiev, IT-entrepreneur, and researcher in Place marketing. One of his current interests is the interrelation between IT-solutions and different city management tasks.  He tries to look at place management concepts from an IT-entrepreneur’s point of view.

Leave a Reply

Your email address will not be published. Required fields are marked *