It is…
The Apple keynote maps feature that was showcased on stage in 2012. Apple broke away from Google Maps for the first time and Flyover mode was touted as a differentiating feature.
I did…
I was brought in to build, scale and then lead a team of about 50 Artists, GIS Editors, and a few Engineers to make sure the cities we were delivering for launch and beyond were up to Apple standards. While my role was clearly the one of Project Manager, once things were rolling and due to the time crunch I would step in and help with any data editing (think of Z-brush), Maya molding, Photoshop work for textures and general Quality Assurance by visually scanning city data.
Deliverables
To meet our typical launch cycle of September, we needed to complete approximately 70 cities in 5 months! This was no small task especially with the very small group of editors based in Sweden.
Results
The details
Maybe you remember but in case you don’t, Apple Maps was launched in 2012. Before Apple Maps, Google Maps was the foundational maps for iOS. One of the keynote features that was presented was Flyover. It was available in select cities around the world at launch and is based off Photogrammetry.
What is photogrammetry?
Photogrammetry is the art and science of extracting 3D information from photographs. The process involves taking overlapping photographs of an object, structure, or space, and converting them into 2D or 3D digital models.
Photogrammetry is often used by surveyors, architects, engineers, and contractors to create topographic maps, meshes, point clouds, or drawings based on the real-world.
Now before the launch of Apple Maps in Fall of 2012, Apple acquired a company based in Linköping, Sweden. This company was the the leader in making these 3D cities and it was my task to embed myself in Sweden for a month to learn their editing pipeline and ultimately rebuild it back in Cupertino. As you will see, things were a little rough and needed quite a bit of curation before going live.
While the previous screens were beautiful examples of the final product, what were were initially getting from our algorithms was not exactly picture perfect.
Texture maps drapped over bridges, lots of holes like these fences, and the Eiffel Tower was not instantly recognizable to say the least.
This is hard! I mean really hard to do especially back in the stone ages of 2012.
Why do you think white is used a lot on roofs?
Of course, it reflects light and most importantly heat. This helps reduce HVAC costs but for our algorithm, white was the most difficult color to understand depth. So you can imagine how many buildings across a city have white roofs and almost all of them covered with spikes.
On the flip side, shadows were troublesome as well. Large areas in NY and Barcelona for example were filled with mountains of “Black Snow” as we called it and needed to be shoveled out.
As mentioned earlier, I spent 3 weeks in Sweden to learn the editing pipeline. This was early summer 2012.
Everyday in Sweden and on the flight back, it was all about documentation and learning how to fix problems.
Back in Cupertino I lead a team of 3 people and we spent the next month interviewing and staffing up a team of editors in a now demolished building where Apple Park is today. Typical with Apple, they really wanted to surprise and delight on launch day and prevent leaks. Because of this we worked in a building that was already vacated for demolition. See that event below, after we moved out of course.
With such a large staff we were building, we had to create training materials and documentation on Wiki pages (remember those). As the team grew we even started tweaking the process to squeeze out efficiencies.
We augmented the process by adding familiar tools like Maya to evaluate the curation and this provided opportunities like adding small details.
Eventually the staff grew to 50 editors and I had daily stand up meetings. It almost felt like a monologue for a late night show but we had a workflow that was running through cities and tons of data at a fast pace.
Of course we kept things organized using off the shelf tools like Jira for bug tracking and task assignments. This was necessary to hit the number of cities expected for launch and at times were running up to 80 house per week.
Even with that effort the maps launch was less than stellar. There were web sites dedicated to collecting a lot of the errors with embarrassing headlines.
Regardless of the initial start out of the blocks, I was very proud of what we were able to build in such a short amount of time. For me, it was literally jumping into a fire and tackling all these issues of large scale production without getting burned. At some points we were up to 80 hours a week which you can only do for so long. Post launch we were able to stabilize and continue production at a far normal pace.