I was invited to participate in a panel discussion on Geospatial Intelligence for #LetsTalkDeepTech Webcast hosted by Swiggy. I talked about the history and evolution of this space and gave a deep dive into solutions for deriving intelligence from imagery.
Below is the a longer version of my talk on evolution of location intelligence with some references. I also share a copy of my presentation at the end. Hope you find it useful. Agree/Disagree with my views? Let me know in the comments.
Evolution of Location Intelligence Technology
What we refer to as Location Intelligence Technology evolved from the disciplines of Geographic Information Systems (GIS) and Remote Sensing. Historically these have been niche specialist areas, where government was the biggest creator of spatial data and the biggest user of these technologies. Spatial data has been used to manage land records, transportation and utility networks, urban planning, census mapping and so forth. People who worked in the field were specialists with years of training and the tooling was primarily desktop based systems. Most data was painstakingly collected in the field, compiled and released after years.
People who have been in the domain distinctly remember when all of this started to change. Around 2005 when Web2.0 technologies were transforming the web, Google Maps was released. It’s innovative use of map tiles in a slippy map interface was easy and intuitive to use. It made the spatial data accessible to users – and most importantly developers. People started putting all sorts of information on top of the map and creating mashups. Google Maps API became the most popular API on the web. This started the transition of the industry towards the web.
The next big step was mobile. As smartphones with GPS chips became mainstream, real-time location was accessible on the device all the time. This enabled not only navigation and local search but a whole range of location based services (LBS) – including things like fitness tracking and location-based games such like Pokemon Go. More importantly, users were now collecting GPS tracks on their phones and taking geotagged photos. Creation of spatial data was getting decentralized with user-generated content and community mapping efforts like OpenStreetMap took off.
Now we are in the world where the amount of spatial data being collected and generated at a massive scale. Ride-sharing companies like Uber and delivery companies such as Swiggy have access to billions of location pings from vehicles and delivery personnels. Connected-cars are capturing data from streets. There is a proliferation of satellite constellations imaging the earth in high-resolution at a cadence that was unthinkable a few years back. This has created unique opportunities to derive insights from it. The tech industry knows this. All large tech companies like Microsoft, Facebook, Apple, Amazon, Uber, Tesla are building their own mapping solutions to solve their business challenges and have large geospatial teams working on location data.
They biggest change has been this shift from static authoritative data towards a feed of large amount of imprecise and noisy data.
The tools that have worked well in the past are not able to work effectively with large volumes of streaming data. While many other industries and scientific disciplines have successfully transitioned to distributed data storage and processing techniques – geospatial has been lagging behind.
But this is changing rapidly. Emergence of data science is expediting convergence of tooling that is able to work with spatial data with the same fluency as other data types. Most databases now support distributed architectures and there are GPU-accelerated spatially enabled databases. Distributed storage and massively parallel processing is now available through systems like GeoMesa and Google Earth Engine.
There are efforts across the industry to solve the technical challenges presented by data volume and accuracy. There are also other important issues such as data privacy that need to be solved. But I am really excited at the direction of the industry and looking forward to seeing how we will solve these problems together.
Location Intelligence From Imagery
I also presented a technical deep dive into solutions for deriving intelligence from overhead and street-level imagery. Below is my presentation with links to external references.