DigitalGlobe and Big Data – The Necessity of a Platform-based Approach

Satellite imagery, content provider and remote sensing spacecraft operator DigitalGlobe is engaging with big data technology, investigating new verticals and planning a new platform-based approach later this year.

By the editors

This will mean they can offer more scalable products and services. John-Isaac Clark, who is Director Product Management Geospatial Big Data at DigitalGlobe, explains all.

Introduction
In the last decade, John-Isaac Clark has worked extensively in the GIS and remote sensing industry. Prior to coming to DigitalGlobe, he was a co-founder of a product and services company based in Northern Virginia which worked extensively with Google’s geospatial products, making them accessible to new user groups as of yet unfamiliar with GIS and location-based technology. His experience in building solutions based on a platform will come in handy in his new function at DigitalGlobe, where he is now Director of Product Management at Geospatial Big Data.

Making things more scalable
In his new function, he is finding ways to integrate evolving technology with products and solutions in highly-scalable ways so that today’s business user can easily leverage it. This evolution is necessary due to the ever-increasing amounts of data that have turned GIS analysts into data wranglers, resulting in them spending more time on preparing data rather than analyzing it.

His solution to this problem is two-fold. The first part deals with creating a platform where everyone has access to DigitalGlobe’s and potentially other people’s data: “The idea is that experts can come and create new algorithms on large sets of data in one place, whereas before it wasn´t possible. Going one step further, we would provide human/machine-learning intersections, where a combination of an algorithmic result can be trained by the human classification mechanisms through crowdsourcing capabilities as- a-service. By doing this, the slew of licensed data can be operated on, or potentially served back to an enterprise in a manner in which they don´t have to manage it. Finally, we want to be able to take insights gained from that data and make them accessible in a way that today´s modern business user is used to seeing everyday.”

By realizing both, many people could potentially take advantage of this improvement, says Clarke:  “I believe that leveraging the building blocks of this platform will enable this to happen. We just have to abstract away all the cool science and awesome capabilities that occur behind the scenes in order to deliver a business tool or tools. We can do this either through our own products or by creating a product by working with industry and vertical experts who already understand this space to deliver insights and capability.”

fiji_north_jan27_2016_dg

The ideal satellite imagery platform
A trend that Clark sees at DigitalGlobe, is to allow answers and insights to be derived where the data resides, which is inside commercial cloud infrastructure. Clark: “Getting access to large data volumes in the cloud is a game-changer, because a cloud-based infrastructure not only eliminates the need for physically moving data, but also the requirement of the traditional investment in hardware and content licensing, the manner in which the industry typically obtains a lot of data from us or other satellite imagery providers.”

This trend necessitates a different approach to address users, as the industry evolution is at a critical point where the amount of data available is far outpacing the capabilities of internal enterprise tools to manage it and extract the type of insights and answers that users want to acquire. Clark says this necessitates a platform-based approach that changes the way in which users access and make use of the data, whether it’s satellite, vector or terrain data. This is why DigitalGlobe has been working on such a platform which will be due for release this year.

To get an idea of what to expect, Clark describes his ideal satellite imagery platform: “The ideal satellite imagery platforms must contain a vast amount of pre-existing content, which immediately gives them value. If you come to this platform, you can have access to 80+ petabytes of content  on day one. The second criteria is about creating an ecosystem, where people can potentially derive new value from others who bring their own algorithms or write new ones. The final criteria is the importance to augment all that with the current capabilities science can offer at the present time and the ability to verify that using human beings, if necessary. The intersection of both machine and human learning together enables higher levels of accuracy.” An example of that final criteria is platform integration with Tomnod, a DigitalGlobe project that uses crowdsourcing to identify objects and places in satellite images.

He has three potential user groups for this platform. Firstly, he discerns a group of developers, such as GIS and remote sensing scientists who develop algorithms or workflows to run against the data to answer their questions. “By making their work public and also potentially licensing them to other users, we are creating a market for them to monetize the work they do.”. Secondly, there are large companies or organizations that have been providing information and insights to support various different verticals and who have a number of proprietary algorithms and capabilities that they use to generate this. Clark: “They come to the platform to run these at scale and across a far vaster collection of data compared to their own infrastructure.” Finally, there are the business end users, who just want to get an answer without necessarily understanding how the technology behind everything works.

fiji_north_feb28_2016_dg

Verticals
Apart from data and technology, subject matter  expertise within specific domains is essential for success when it comes to big data, which leads us to the question of what verticals are of interest to DigitalGlobe currently, and which will be the new and emerging ones.

Right now, the company is working with developers to build capabilities for financial markets and generating information for insurance businesses. They will be used, for example, in cases such as disaster impact, but also for construction, transportation, supply chain logistics, space and telecommunication. Clark: “of course there´s also energy, where even despite the big depression in oil and gas prices there are numerous use cases where being able to operate on this data at scale can provide insights that can save money.”

A completely emerging new space is that of UAS. The establishment of rules by the FAA enable commercial routine use of drones for many things, and a combination of terrain, vector and satellite data can help drones to fly around in major urban areas without crashing. A solution for this is offered by DigitalGlobe’s partner PrecisionHawk, called LATAS, which stands for Low Altitude Tracking and Avoidance System. This is an onboard system that provides flight planning, tracking and avoidance for every drone in the sky using real-time flight data transmission based on existing world-wide cellular networks.

Another new area are microsats; offering new capabilities and business opportunities. Clarks sees these as a valuable addition and partner, but thinks it´s important to understand what they can and cannot do: “It´s the combination of the things working together that excites us. We’re looking forward to bringing better solutions to the market jointly as these other providers get their capabilities into orbit. Our hope over time is to bring DigitalGlobe data and potentially multiple other micro satellite providers all into one place. We believe that the greatest value for the industry will be created by enabling various algorithms and capabilities to run on all the data.”

Clark describes the process of engaging with verticals, industries and markets and determining what works and what doesn´t as a rapid, interactive and dynamic process. It´s also not uncommon for a use case to span multiple verticals. In this instance, the idea is to enable a number of people at scale quicker rather than going into single vertical. It´s important to move quickly, says Clark: “we believe there´s a greenfield opportunity for our platform, but it´s really important that we act quickly by finding and delivering the value. If we can´t provide a solution for a problem yet, it´s important to understand what it would take to do that and then work on adding new partners with different capabilities to achieve it.”

For more information on DigitalGlobe’s Geospatial Big Data Platform (GBDX), have a look at www.geobigdata.io

www.digitalglobe.com

 

 

 

Share on facebook
Facebook
Share on twitter
Twitter
Share on linkedin
LinkedIn

Related posts

Scroll to Top