Google maps is one of the most commonly used online services in the world. Most would agree, where there is google, there is a way. It’s mind-boggling how accurate google maps is in most parts of the world. While all of us love the service, very few actually know how it really works. How did google collect so much data about the world? How do they constantly update the maps? Traffic? Speed limits? Business hours? Reviews? As a curious PM, I dig deeper to unravel the mystery behind some of these questions.
Why did Google launch maps?
The answer is closely related to the company’s mission. Google is first and foremost a data company with the mission of organizing the world’s information. Documenting the entire world and making it easy to navigate aligns well with Google’s mission. Google first set out to create a catalog of where businesses were in the world and soon realized that there needed to be a useful and meaningful map of the whole world.
Google set its foot in the maps industry by acquiring Where 2, a company with groundbreaking technology that allowed navigating to different parts of the map without needing to refresh.
What data do you need for maps?
Today, google has perfected maps using over 20 Petabytes of data. They have a ton of sources that help them collect this data, but at a high-level the three most important ones are Satellite, Government and People.
Satellite and Street View
Satellite and Street view has helped google identify where places are around the world. It has also been a crucial part of keeping the ever evolving parts of the world updated. Using satellites, google can stitch together pieces of the world and street view helps them enhance the accuracy of the maps. Google not only knows roadways, buildings and addresses, but street views also help them isolate speed limits of each roadway, business names, traffic signs, direction of traffic, etc. Street view is now being used for navigating inside structures like airports and malls. As of 2019, street view alone has captured 170 billion images.
People often believe that Google has their own satellites. However, that is not true. Google uses satellites from agencies around the world. If you look closely to the footer of google maps, you will find they often credit the agency or the satellite that was used to capture that part of the planet. Google however does use a small army of people to capture street view. It is not too rare to see a google car capturing 360 degree views of its surroundings while driving at 6o miles per hour. For busy streets or mountain peaks, there is Street View trekker.
Government and Third Party Sources
Google layers in data provided by the government of various countries and other third party sources like United States Geological Survey and the National Institute of Statistics and Geography. These agencies provide google with the most comprehensive and authoritative data. Using the Map partner programs and tools developed by google, agencies can submit detailed vector maps that will help google demarcate boundaries, identify bike paths, new houses, etc.
People help google with bringing all the pieces together in two forms:
- Employees – Around the world google has data operations teams that help collect and verify maps submitted by agencies, vet data aggregated from street view, go through feedback reported by users and most importantly, develop algorithms to improve the accuracy and usability of maps.
- Crowdsourcing – Google has various tools that can be used by users to provide their feedback
- Localguides – Collect photos, reviews, hours of operation and other subjective data on local business. That’s how google can accurately show us the details for each business.
- Edit Maps – Users are allowed to make personal contributions to the maps by editing every part of the map which is open to public reviews.
- Location Services – Mobile phone GPS is used to accurately predict road traffic, crowd at restaurants and metros, and most importantly, to navigate real-time on the maps. Location services are also used to improve the quality of maps. If google is recommending roads that most people don’t end up taking, google can assume that the road is not accessible anymore. They will use this data to divert traffic to alternate routes till they can update the maps.
- Report/Feedback – Like most softwares, maps is not perfect. It banks on people to report issues and provide feedback on the maps. Google uses these reports to improve the maps on a daily basis.
Bringing it all together
Google invests millions of dollars into developing machine learning algorithms that can stitch the data together and make it meaningful. Computer vision is used to identify tons of information from a single image that has been captured by satellite or street view. GPS location is aggregated to predict traffic and popular times at a restaurant. Google has now developed algorithms to accurately map 3D models of building structures on the map.
Machine learning is also used to drive revenue. If you frequent a local restaurant, then google understands your choice and serves you relevant ads. They promote relevant businesses on the maps itself by showing logos in a subtle form. There are tons of other ways that google uses maps data to drive revenue.
Google has collected so much data through maps that the possibility of innovation is limitless. Some of the newly launched and upcoming features are: