Esri UC 2017 Plenary Report

We look forward to the Esri UC Plenary session each year, as it always profers an important vision, but also new insights, thoughts and surprising technologies.

This year’s conference slogan was “the Science of Where.” This was the largest conference in 30 years with attendance at approximately 16,000 people. As always, Esri founder and president, Jack Dangermond, had everyone get up and greet the persons sitting next to them at the Plenary, and encouraged everyone to continue to do this throughout the conference.

This time-tested way of getting to know people is an ice breaker to help people come together and share their GIS stories, and perhaps form bonds that will lead to more wonderful work together.

Dangermond talked about how from the very beginning GIS was applied to science of all kinds. “But it wasn’t until 1993 when Michael Goodchild wrote a paper called GI Science, it changed the game,” he said. “The world began to realize that GIS is a science in its own right and incorporates other sciences, like geography, computer science, decision support science, and incorporates them like a meta science. What is the Science of Where? Simply stated, it’s the science of geography and technology of GIS, this is a science framework for applying science almost everywhere.”

To meet a challenging, changing world, GIS is a fundamental language and framework wherein we can organize our maps and data and analyze shapes and patterns, and can do design and planning, make decisions and take actions.

“Why is this so important right now? We’re living in a world that is increasingly challenged, with growing population, issues of climate change, loss of nature, social conflict. Organizations are being confronted with changing world, we have to do everything we possibly can to better understand, and to form better collaboration, to address these types of ways UPS made their tremendous advancements.”

Jack said that the Web GIS is the modern GIS architecture that uses Web Services to integrate teams and individuals, departments and organizations with shared knowledge and collaboration and to improve efficiency.

All types of data can be used with Web GIS, brought together using web maps, web scenes and web layers. Smart mapping makes use of computation and logic to create maps automatically and use exploratory data analysis for a result of beautiful maps and greater understanding.

People are now integrating science into the design process, so they can evaluate consequences of design options. Web communities and web GIS enable community engagement, by organizing and managing all the community interactions. An example is a hub in Los Angeles, where citizens want more information about their city, and the leaders want to hear from citizens. They’ve been working on policy initiatives that communicate with citizens and seek their feedback.

A whole new scale of GIS is available, that interconnects separate systems with web services and creates systems of systems. An example is the city of Las Vegas, where the water utility is now sharing their information with utilities, and doing it as web services. This keeps the utilities from digging where they shouldn’t. This is a great example of “the Science of Where”, or in this case, “Where Not.”

ArcGIS Capabilities Provide a Complete GIS Platform

 Flagship GIS product ArcGIS is considered a complete integrated GIS platform, and has been engineered as an open platform, full services-based and distributed. It’s available on your own machines or on the cloud, and available for individual and teams of organizations.

The platform is a huge engineering feat, said Dangermond. An integrated complete platform, ArcGIS organizes and manages all aspects of a GIS, it also includes the information products created from the data. The Portal is used as technology to provide a common integrated user experience to bring different items together, with different parts, such as server, desktop, enterprise online. This “distributed pattern (new) systems of systems” is about integrating other patterns to connect everything.

ArcGIS Content is a fundamental part of the platform, that has thousands of ready to use maps and datasets from Esri, and includes base maps, soils, imagery and demographics, boundaries, and a large amount of topical information sets.

Essentially this is a living atlas, “the foremost collection of global information on the planet.” The model is content as services.

A core part is data management and compilation that provides new workflows and tools – improved editing, improved geocoding and location finding, and CAD integration.

In the area of Field GIS, in addition to Survey123, Collector, Navigator and Workforce, this year Esri released a new version of ArcGIS Online, so you can take maps to the field and be able to do markups, in a connected and disconnected environment.

Smart mapping tools are improving not just with 2D cartography but 3D thematic representations, and fast display. In the Pro space they have added new cartography tools, such as SVG support, dynamic charts in layouts, measured grids, multi-scale drawing, production charting and mapping progressing.

In addition, there are new vector tiles for speeding up your display, that can project your own vector tiles in your own projection. In both 2d and 3D environments.

Progress has been made in integration with Adobe Creative Cloud in both Illustrator and Photoshop, so graphic display users can use the two products together.

In the 3D space GIS is now viewed as a common 3D platform for 3D visualization analytics, with its ability to bring in data sources of all types, BIM and CAD drawings into GIS from LiDAR, and visualization from LiDAR into GIS, extending the language of maps into the language of 3D GIS work. More tools are now available for urban planning and geodesign as well as for navigation and Pro for visualization. The News Scene Viewer 2D and 3D visualization are very powerful. New data types are available, now supporting 3D objects and meshes with specification of IPS, and innovations of Augmented Reality and Virtual Reality.

As a complete imagery platform, ArcGIS had three major components: 1)  the support for all leading sensors –  big satellites, small sats, air photos, multidimensional data, radar, drones 2) dynamic image processing, that means image processing on the fly, which means if you have an image and you pan and zoom around on it, it does dynamic processing instead of processing the thing and waiting until it is done and then using it.  It looks at raw imagery coming off the sensor and can dynamically process and analyze all of it. 3) Classification or NDVI change analysis, new things with Image space vs. geographic space. The photo interpreter wants to see data in image space and dance back and forth between geographic space and image space, and can capture from a warped image and digitize and view it in Esri’s database.

This dynamic image processing capability makes it possible to put software right next to large massive image stores of dozens of large data sets in the cloud, and process imagery on request, dynamic processing next to data, and server serves out image services and pictures of results in a second.

This is already being done with Landsat that has coverage for the whole world, and this can be stood up in the cloud within a few hours. In the fall Esri will support Sentinel 2, the European satellite, and will release global elevation data for the entire world.

New progress has been made in the area of spatial analysis, web-based analytics, new tools for spatial statistics, being able to do vector analytics in “space time cubes,” noted Dangermond.

The integration with science through the R integration and ArcGIS API of Python allows Improvement Processing. “We’re using the parallel processing capabilities of the new GPU to parallelize the computation like on a desktop for raster analytics, which improves speed. And models as a service lets me take a cool model I built, and right click to send it to my enterprise server, immediately it comes alive with a service that others can use. This is going to change how we share our tradecraft in the spatial analysis world.”

A new product announced earlier this year is Insights, that provides a whole new way to do spatial analysis in an intuitive and visual way, what academics call “exploratory data analysis and visualization.” Any type of database can be explored in terms of space time patterns, charts, graphics, etc., and can be shared with colleagues in other fields.

For big data, Esri has been working on tools to make spatial analytics faster and massively scalable out to tens of millions of observations or more. Using the Geoanalytic Server, the space time work previously only done on the desktop can be done at a performance of ten to a hundredfold now, with tens of thousands of images done at scale and almost immediately delivered.

Real time analytics has taken a jump to be able to make hundreds of thousands of observations a second and analyze high velocity data streams, monitoring and alerting, integrating sensor networks and the IoT – in real time. This is done using GeoEvent Server in situation rooms and other places requiring the use of real time analytics of big data.

ArcGIS has been designed as an open system and platform which is standards compliant. This is accomplished in three ways:

• by supporting leading open standards and formats, OGC and others, compliance with certifications
• by directly integrating Esri’s products with other leading COTS technology which bypasses the need to spend a lot for system integration; they are doing this with Autodesk right now.
• ArcGIS designed to be open, open access to data, an extendable designed architecture that can be embedded.

The topics above were covered in the afternoon Plenary in greater detail.

Artificial Intelligence for Conservation

A big announcement made at the conference was the collaboration between Esri and Microsoft AI to provide greater access to artificial intelligence and other advanced technologies in land cover mapping through a newly aligned grant process.

Lucas N. Joppa PhD, chief environmental scientist, Microsoft AI, spoke about how they have been working with Esri and the Chesapeake Conservancy, to determine the value of joining forces to use artificial intelligence to both accelerate conservancy work and empower more organizations in more places to sustainably manage their lands.

Land cover mapping is a critical part of conservation planning, but current methods for collecting detailed geographic data are highly labor-intensive.

“We used the original Chesapeake land cover map as training data for an advanced AI approach called deep learning to create a single land cover classification algorithm with potential to work across the US,” said Joppa.

The early results depicted that algorithm integrated into ArcGIS, a location in the Chesapeake watershed, and four types of information, with 1 meter resolution NAIP imagery from the Living Atlas, the original land cover map, the map that’s being produced by their deep learning algorithm on the fly classifying water, fields and impervious surfaces like roads and houses. And in another map, the mixed region showing it as a mixture of probabilities from their model across all of those categories.

“It lets us see how the algorithm is thinking and what we might need to do to improve it,” said Joppa. “We can use this same algorithm to classify places that it’s never seen before. By using our algorithm almost immediately we can update our map with more current information.”

“Our algorithm has a lot of learning to do about features like roads and houses. We can use those features as new training data for our model, and as we get more data from more places it will be eventually able to classify land cover from across the U.S.”

The classified output from the algorithm at scale for all of Oakland County, Michigan (a case study shown during the Plenary by Phil Bertolini, Deputy County Executive/CIO Oakland County, Mich.) was produced in just a few hours using Image Server in the cloud. This is an important benchmark, as for the first time, the creation of these high-resolution land cover maps will be able to keep pace with the increase in production of high resolution imagery.

“AI done well disappears into the background and lets users concentrate on task at hand,” said Joppa.

This is still in the research phase and is only one way Microsoft has been working with ESRI in the past year. “We’ve been working to align our technologies, and as well as our resource granting programs,” said Joppa. “Eligible organizations will have access not only to Esri’s ArcGIS software but also Microsoft’s leading AI tools like our cognitive tool kit and our Data Science Virtual Machine, which hosts popular apps like R Studio and Jupiter Notebooks. Our relationship with the Conservancy started with a grant just like the one we’re doing today.”

The aligned grant process will allow eligible organizations and individuals to receive Esri’s premier geographic information system (GIS) software as well as access Microsoft’s leading cloud and AI tools. These include Esri’s ArcGIS and Microsoft’s Data Science Virtual Machine and Cognitive Services offering, in addition to Azure credits.

“Spatial analytics is a crucial part of conservation, because it is an inherently geographic field,” said Dangermond. “Understanding environmental data in the context of location is necessary when making decisions about protecting fragile ecosystems. The ability to analyze the vast quantities of data present in our environment is critical, and this collaboration allows conservationists to do their job more accurately than ever before on a wider scale.”

This brings us back full circle to the “Science of Where.” It’s not as though we haven’t heard of the importance of location, location, location before, but this research brings us to an important juncture, where technology can assess location that it has not previously seen, allowing the analysis to take on an expanded role in the protection and understanding of our environment.

Tags: ArcGIS, artificial intelligence, cloud, crowdsourcing, data, ESRI, geospatial, GIS, imagery, indoor mapping, intelligence, mapping

Categories: analytics, ArcGIS, ArcGIS Earth, ArcGIS Online, climate change, cloud, cloud network analytics, drones, field GIS, geocoding, geospatial, GIS, government, GPS, location based services, location intelligence, mapping, OGC, satellite imagery, smartphones, UAVs, utilities

This entry was posted on Thursday, July 13th, 2017 at 7:24 am. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.