ADDING TIME TRAVEL TO BING MAPS

Background.

If you’re familiar with Google ‘Street View’ or Bing Maps ‘Street Side’, you know that this imagery allows map users to engage at human-scale to virtually ‘walk’ around the streets of a city or see the front facade of a house or other destination. In 2010, Microsoft researchers created a novel computer vision-based augmented reality algorithm that performed alignment of still photographs to a set of base-layer panorama images. The technology seemed like a perfect application for Bing Maps Street Side and was tested using crowdsourced images from Flickr. The researchers were able to skew, align, and embed the Flickr photographs within Street Side.

My Role.

The prototype was buggy and not packaged into something consumable or lovable for Bing users. That’s where I came in. As a Program Manager in Bing, I was tasked with leading the team to create a “Community Technology Preview” that would engage both internal and external customers and maximize press coverage for Microsoft. Our lean team consisted of one product manager (me), one researcher, one front-end engineer, and one shared designer. To hit our marketing dates, we had just two months to make a polished demo ready for public distribution.

Challenges.

The biggest hurdle was designing the UX. The goal was to make the experience rich and satisfying, even if brief. We created a simple list view that would allow users to jump from image to image, so that users didn’t have to manually and cumbersomely ‘walk’ the streetside imagery. Some neighborhoods had a dense amount of images while others had very few and both of these spectrums presented opportunities for users to get stuck. To address these UX traps, we showed the next closest image in the list no matter how far away, while also giving the user an option to quickly transport themselves to another neighborhood by clicking on “explore a new place”. Another consideration was abiding by copyright licenses. We only embedded photographs that were sourced from users who had uploaded them to Flickr and tagged them with Creative Commons licenses, giving explicit permission for 3rd party use. On a daily basis, we checked each image license and removed content where licenses had changed.

Highlighting the Value.

Though we had only a short few weeks to deliver something, we wanted to pack a punch and build something worth telling our friends about. It was delightful to see many of the photos superimposed within the Bing Maps imagery; however, most of the images were not particularly compelling. The ones that stood out to us offered a significantly different perspective from the underlying street side imagery. In order to showcase the best content and ensure users would see them, we curated the best examples and featured them on the app homepage where they would be easily accessible.

By highlighting the best crowdsourced images, a user could instantly gain a new appreciation for the culture of a city including the people and traditions that bring city streets to life. The experience also provided the user with an appreciation for the changes in seasons, the beauty of a scene when it lights up at night, and the history and artwork that had since been lost. There were snapshots of murals that had since been painted over, all in their original context and splendor. These images would otherwise have been disjointed from their original locations. It was the context that gave the images more meaning and enhanced their value.

Though impossible to convey through the still screenshots shown here, we also spent time perfecting the transitions between images. The animation and the unique skewing of the images to align them with the underlying streetside imagery enhanced the innovative feel of the experience, which ultimately led to more interest and press.

Time Travel.

A few weeks into the project, I reached out to city archives in Seattle and Vancouver and was granted permission to utilize their database of historic photographs including photographs that were over a century old. The time consuming part was to match the historic images with places that still exist today. While the previous crowdsourced images had been automatically geo-positioned into Bing Maps, many of these old photographs looked so different from the original locations that our algorithms couldn’t match them to a location, or at least without rough GPS metadata. I stayed at work late each night virtually wandering the streets of Seattle and Vancouver within Bing Maps to see if I could match by eye the locations of the photographs. Some of my favorite examples that we were able to add to the experience showed buildings under construction, horse and carriages on dirt roads, people in the streets from 1898 and 1948, and a long-gone Japanese Archway on Hastings in Vancouver photographed in 1901. The historic images became wonderous easter eggs in our experience (see the gallery at the bottom of the page for more visuals).

This animation captures Seattle’s Pike Place Market as it looked over a century ago.

The Results.

It’s not often that research ideas make it into production, which is part of what makes this project particularly special. Shortly after launch, our app called ‘Street Shots’ was featured on a popular TED talk given by Bing Maps Architect Blaise Aguera y Arcas that quickly gained over a million views (see minute 3:00). This was an incredible moment for the team and is still one of the highlights of my career! The project received a great deal of press attention as we met all of our goals.

It was our ability to work in an agile manner across the organization that allowed for us to utilize a novel technology and make it consumable. We started with a test of a raw algorithm and we were able to rapidly ship something beautiful, captivating, and unexpected. Having the space to be creative was incredibly motivating and there was a real sense of camaraderie on the team. While it’s been over decade since I worked on this project, I still remember it as one of my all-time favorites.

GALLERY