Friday, 17 August 2012

How would an algorithm data mine Freo's visual identity?

There is an interesting article on Atlantic Cities about data mining a city’s visual identity. It got me thinking about Freo’s visual identity.

It is generally accepted that there is no city quite like Paris.  For me, the allure of Paris isn’t so much its great monuments and awesome museums and art galleries. What I'm fascinated by is that I can take a stroll down any residential street in one of the twenty arrondisements and there is no escaping the fact that I’m in Paris. I agree with
Emily Badger, the author of the article, who opens with her view that Paris looks like, well, Paris and like nowhere else on earth.

(Upon editing this article, my wife started chuckling and prodding my tummy. In between jiggling my man boobs, she also claimed that a big part of why I like Paris so much has a lot to do with bistros, butter and pastries. I don’t know what she is talking about.)

Cool algorithm that data mines the visual identity of cities

Emily points out that people who have been to Paris become adept at identifying the place. In a recent study, researchers showed participants a sampling of images of Paris as well as decoys from eleven other cities. The results showed that the participants nailed Paris 79 percent of the time. (Check out the game here.)

The study allowed the researchers to conclude that people are remarkably sensitive to the geographically informative features within the visual environment. I reckon that this statement characterises the majority of us Freoites.

The findings of the study has led to an algorithm being developed that attempts to data mine the visual identity of cities. Below are some results from Emily’s article:

Random and Extracted visual elements from Paris via
Random and Extracted visual elements from Boston via
Random and Extracted visual elements from San Francisco via

Interesting stuff. The researchers found that the algorithm had problems data mining the visual identity of American cities. They concluded that this may have been because many American cities are not all that unique. 

So what does this mean for Freo?

1. The Local Identity & Design Code: Central Fremantle

Freo’s visual identity has kind of been mapped. A couple of years ago, ratepayers funded a study of Fremantle’s identity. This study produced a very, very, very comprehensive Local Identity & Design Code: Central Fremantle document.

If we all agree that Freoites are remarkably sensitive to Freo’s geographically informative features within the visual environment, then what do we think about this document?

And how is the City using it to inform the design guidelines that will support Scheme Amendment 49?

(Click here to check out this document.)

2. A problem with car and road dominance

The researchers discovered that one of the main reasons that American cities lacked uniqueness was the car. In fact, they found that the prevalence of car brands and road features contributed to the overall blandness of these cities.

This is something that we should all be aware of when planning for the future of Freo. I’d hate to think that future algorithms that data mine Freo’s visual identity come up with cars brands and road features rather than awesome West End architecture.

3. Freo’s a lot like Paris

To paraphrase Emily Badger: compared to Perth, Freo looks like, well, Freo and like nowhere else. That is really important! (Note to wife: first exclamation point in general text of one of my posts for a really long time.)

I firmly believe that Freo’s redevelopment doesn’t have to be a bad thing.

But it will be crucial that new development is built up to a standard and not down to a price. I want to be able to walk through Scheme Amendment 49 territory and feel like I’m walking in Freo and not a clone town. That means having an idea of Freo’s visual identity and insisting that these elements are reinforced.

No comments:

Post a Comment