Back in 2024 media artist Yufeng Zhao fed more than 8 million Google Maps Street View panoramas into a machine learning model that identifies and transcribes visible text. The result was All Text in NYC – a searchable archive of 138 million snippets of urban lettering gathered from graffiti, storefronts, bumper stickers, menus, flyers, billboards, and more. A search engine that turned the whole of New York City into a giant text document you could explore by keyword.
His latest project, NYC Street Clock, feels like a natural continuation of that idea. Instead of turning the city into readable language, it turns the city into time itself.
The concept is beautifully simple. Visit the site and you’re greeted with the current time displayed entirely through numbers found in Street View images captured around New York. Every digit is sourced from somewhere in the city: a bus route, a deli price sign, an apartment number, a lottery ad, a parking meter, a glowing storefront clock. The hours and minutes continuously refresh, as time passes, displayed using fragments of urban life found on Google Maps.
Just in case Yufeng Zhao ever reads this blog, can I suggest an idea for the next project: NYC Typewriter – an app that allows users to type in a sentence and then visualizes it using words pulled from the All Text in NYC database.. It would let people create their own “ransom note” style messages – like the one I created at the top of this post.
Via: Webcurios
