Expert Roundup: How is Mapping Innovation Playing Out? Part I
Mapping is one of those foundational “meat and potatoes” topics in Street Fight’s repertoire that buttresses local commerce. But despite its longstanding positioning at the center of that world, and its mature status, it still somehow continues to show rapid transformation and innovation cycles.
Local search and mapping are also more important than ever. As businesses reopen, they’ll have to communicate dynamic business details such as mask policies and other variables. These details will have to be a core component of the local search and mapping experiences — at least for businesses that do it right.
So, to define the current state and future trajectory of mapping, we’ve rounded up top industry voices and thought leaders. This is part of Street Fight’s new monthly ritual in which we tap our community to provide insights on each month’s editorial theme.
Starting here in Part I, below are the insights we were able to gather from the community.
Giovanni Lanfranchi, CTO, HERE Technologies, on Mappint’s AI Revolution
Mapping technology is going through a revolution driven by AI. Our grand challenge at HERE Technologies is to design AI-based mapping technology that helps us understand how the world moves and interacts. The real world is complex, infrastructure varies, and the driving and mobility habits within each city and market around the world are unique. So, if you’re going to guide people through a city or train a car to drive itself, there are so many dynamic variables to consider: traffic flow, hazards, roadwork, bike lanes, available electric charging points, time of day, weather, and so on.
Fortunately, AI-based mapping systems are becoming increasingly adept at crunching through all this data. For example, we’re deploying neural networks and deep learning perception models to fully automate the processing and fusion of data from different sources, such as vehicle cameras and sensors, smartphones, LiDAR, probe data, and aerial imagery. As much of the data flows in in real time, you can give everyone a rich and current 3D view of what’s happening around them. At the same time, AI systems learn how cities ebb and flow and can make predictions about future states. This is where mapping technology is going, and it will be important in all kinds of ways – safer autonomous traffic, more efficient logistics, directing first responders to where needed, people flow, and so on.
Beyond this, there are some other interesting problems we are looking to solve that use AI.
One area of research involves doing away with the map altogether. We design maps for machines, but what if a machine has no map? How can it find its way? This is an exciting area of research that involves the creation of neural models that can interact with and navigate through an environment without any explicit prior information about it.
Another area of research involves something called ‘signatures.’ We use an object-detection neural network to create a signature of an object, such as a house or a building. Signatures look like barcodes, and the signature of every building, set of buildings, or street is different in small ways. You can point your camera at a building, and the Signature AI will infer your location. This technology has some interesting potential applications. You could take a picture for a drone to tell it where you want it to deliver a package. You could use signatures to show a delivery driver where to leave a package by highlighting the building in their display or AR view. Or you could use signatures for machinery and goods monitoring in a port, warehouse, or construction site – the signature changes if a known object moves or disappears.
One of the most fascinating problems we’re working on at the moment is how you can relate all the different elements in the map – people, places, objects – to one another. We call this a location graph. It’s a semantic model of the world that attaches meaning to spatial objects, such as “this is a home,” “this is a gas station,” “this is the entrance to the metro.” The graph helps with seemingly simple, yet actually complex queries like: “Find me a fast-charging station along my route where I can also grab a coffee.” We believe this technology will be vital for orchestrating transportation services and fleets. But it will also be useful for brands, gaming studios, and high-footfall venues like malls and stadiums to create exciting location-contextual services, immersive AR experiences, and as-it-happens ad campaigns.
Dan Silver, SVP, Marketing, GroundTruth, on Mapping’s Many Faces
With the wealth of location data available, digital marketers can define areas where their customers go, down to the store level. This level of data gives marketers running campaigns driven by mapping technologies a new level of granularity.
Available technologies can identify precise location boundaries that capture the world we live in and how we live in it. By being able to accurately identify brands and multi-layer points-of-interest through mapping, marketers can fully understand who consumers are and how they engage with their surroundings. There’s a lot to be excited about when it comes to how marketers and brands are leveraging mapping to help define their strategies, including:
Out-of-home (OOH) Advertising: The adoption of new measuring techniques and mapping technology is empowering OOH to keep pace with advertising on digital channels like Facebook and Snapchat. OOH advertisers can better understand who is viewing their ads, which helps them target content more precisely than ever before and better measure the ads’ impact.
Curbside Pickup: During the pandemic, curbside pickup became an essential purchasing option for consumers. Being able to target consumers who buy online and pick up outside a store location has become a new avenue for marketers to reach shoppers. By mapping curbside pick-up zones, brands can now build accurate custom audience segments comprised of these specific shoppers and send targeted incentives and offerings based on their shopping behavior.
Trade Areas: Mapping technologies can help marketers reach mobile users in areas with a high visitation affinity to a brand or store by coupling visitation patterns and audience segments with custom geographical boundaries.
Fundraising: Mapping technologies allow businesses and nonprofits to set the perimeter of their choice all the way down to a neighborhood or even a building. This can be an extremely helpful fundraising tool targeting an event or location for making donations.
Custom Audiences: Mapping helps form the basis of audience profiles. When advertisers want to target a specific group of consumers, the location behaviors are defined by having visited one or more locations. With access to precise mapping and location targeting, brands can rest assured there will be minimal wasted impressions on their advertisements.
Our industry’s biggest challenge is ensuring accuracy and pinpointing exactly where consumers are and where they are going. Using data to understand the local context of consumers is crucial for relevant and effective digital advertising campaigns, and it is precise mapping technology that drives these campaigns to be successful.
Auren Hoffman, CEO and Founder, SafeGraph, On Mapping Tech Stacks
The biggest trend happening in mapping is the proliferation of companies that are rolling their own stack rather than relying on one unified solution (like Google Maps).
Mapping and geospatial analysis have been democratized in recent years, and this trend is rapidly accelerating. Traditionally, many organizations relied 100% on Google Maps for their entire mapping tech stack, everything from the basemap to the underlying points of interest (POI) data. The minority who did not, and instead used a combination of more sophisticated tools, had dedicated GIS departments with advanced mapping capabilities.
As more people with roles outside of traditional GIS become empowered with geospatial data and analytics, companies are now expanding their mapping tech stacks to include the best of class for their specific use cases. Consequently, tech stacks are now diversified to include different providers and platforms for the underlying data, display, and mapping itself.
Jefferies, the investment bank, is a great example of a company with a diversified mapping tech stack that has evolved significantly in the past few years. They are able to layer their own financial data with third-party datasets and conduct complex spatial analysis in CARTO before developing shareable mapping visualizations. The combination of these different providers is what makes Jefferies able to improve the accuracy and efficiency of their equity research.
Replica, which is focused on the built environment, has leveraged third-party data sources and Mapbox mapping technology to bridge the divide between raw spatial data and impactful data visualization, allowing customers to analyze and predict how people move about urban landscapes. With Replica, public and private organizations can glean mobility and economic insights from multiple angles, whether at the national scale or all the way down to the census tract.
Every organization’s mapping needs are different, so the ability to incorporate datasets and analytics tools specific to those needs is a critical part of these evolving tech stacks. For example, some businesses may only need to display POI data on their maps, while others may need to also show street networks in relation to those POIs. The vast array of tools now available — and most importantly, usable for the general public — has made it possible for these tech stacks to be highly customizable to specific company needs.
Stay tuned for Part II of this series next week.