Summary Bullets:
• AWS has launched the preview release of new geospatial tools that will help SageMaker to compete in the market of geospatial and AI applications.
• Enterprises can now access and share geospatial data in more user-friendly formats, such as through APIs. This means that developers without a technical background in geospatial data can deploy applications in different industries.
AWS recently launched the preview release of Amazon SageMaker‘s new geospatial capabilities. The announcement is significant because it is the first time that the cloud computing giant adds geospatial tools to its ML platform, although last year saw the release of Amazon Location Service. That solution already helped developers add location functionality to their applications, visualize maps, search points of interest, optimize delivery routes, track assets, and use geofencing to detect entry and exit events in a defined geographical boundary.
With this announcement, AWS introduces a complete toolset to make predictions using satellite and location data. Geospatial data has many uses in industries including urban planning, retail, automotive, insurance, agriculture, and more. For example, maximizing harvesting of crops in agriculture, predicting car accident hotspots in urban development, combining maps with competitive intelligence to optimize store locations in retail, detecting old gas pipelines in utilities, etc.
In fast-growing sectors such as environmental services, applying AI to geospatial data helps to analyze the damage from extreme weather events linked to climate change. It not only helps in real-time mitigation of the effects of such events, but also to forecast and help prevent their consequences before they happen.
However, leveraging this type of data entails working with massive, unstructured data sets, from multiple sources and in different formats. The entire process, from data enrichment to visualization, can take months. As a result, data scientists sometimes spend a long time preparing and labelling the data, before even getting to write a single line of code.
AWS claims that SageMaker can accelerate and simplify the process of creating geospatial ML predictions by enabling customers to enrich their datasets, train geospatial models, and visualize the results in hours instead of months. To accelerate the task of building models for the most common uses, the company incorporated built-in pre-trained deep neural network (DNN) models, and geospatial operators that help users access and prepare large geospatial datasets, alongside maps to visualize predictions. The process is automated to allow access to geospatial data sources from AWS (such as Amazon Location Service), open-source datasets (Amazon Open Data), or third-party providers (like Planet Labs).
For example, the Road Extraction visualization tool leverages satellite imagery to give users full visibility of what is happening on the ground, in the event of floodings or other emergencies, so that they can send directions to first-aid teams to identify which roads are safe. This can be combined with apps such as Foursquare to see the location of the nearest hospital within the roads which are not flooded, directing medical staff and evacuating people out of the danger zone.
This release is a timely move by AWS, and could be seen as a response to competitive pressure from the likes of Google Cloud, which dominates the market thanks in part to its conspicuous mapping tools. However, other cloud providers including Oracle and especially Azure also offer compelling solutions. Microsoft has Azure Maps, a suite of geospatial services to incorporate location-based data into web and mobile solutions.
Overall, converting location data into knowledge is notoriously complex, and the providers that can offer tools for industry-specific use cases through simplified, automated processes will win the race. The introduction of pre-built ML models will no doubt generate interest.
IT Connection 