Turning Big Data into Big Business - 04/07/2017
Professor Ian Dowman of University College London reports from an event that heard numerous examples of how Earth observation data is being used daily to inform decision making by policy makers, governments, insurers and security providers.
Engage is the latest in a series of conferences organised by DigitalGlobe, the company operating the Worldview constellation of satellites and providing services based on the data they capture. A key component of the business is the archive which goes back to 1999 and contains 90 petabytes of data.
The latest conference was billed as a two-way exchange between data supplier and user and had the theme of ‘turning big data into big business’. EO data was just one of the data sources discussed. There was an emphasis on end-to-end systems involving data acquisition and geospatial analytics using machine learning and artificial intelligence, with many references to big data, the cloud, and crowd sourcing. The presentations showed that the range of disciplines using geospatial data is very wide, encompassing transport, crime, health, insurance, management of refugees and others, all with the theme of data for decision making.
The conference covered developments in Earth observation with the main trends being the development of small satellites operating in constellations. The principal company is Planet which operates a constellation of cubesats collecting images with around 3m resolution. Clear trends to come out from the speakers were the maturity of software with positional accuracy being high and DEMs being widely used.
Move to Geospatial Analytics
Companies like DigitalGlobe have moved from being data providers to being service providers, offering geospatial analytics which use artificial intelligence to provide information to users. DigitalGlobe puts great store on its archive of images that offers multiple coverage over time of almost the whole globe. Within government organisations the European Sentinel satellites collect 8 terrabytes of data a day, which is used to generate products for the European Copernicus programme. The Copernicus services are aimed at policymakers and public authorities who need the information to develop environmental legislation and policies or to take critical decisions in the event of an emergency, such as a natural disaster or a humanitarian crisis. This reflects a growing application of Earth observation to serve the public good and includes job creation through new companies offering value added services.
The main user of EO data is the defence sector, but a number of interesting applications were discussed. The insurance industry is using geospatial data for risk assessment, 3D data is becoming useful in order to know which floor of a tall building an asset is on or whether an object is on the sea or in the air. A presentation from the Bill and Melinda Gates Foundation showed how it is essential that location of settlements is known to ensure that Polio vaccination programmes, which need to cover everybody, can be complete, and also that access routes are well defined. The same presentation showed that settlements destroyed by Boko Haram should be identified so that teams do not waste time and take risks in visiting them.
Other applications include mapping the incidence of crime and identifying routes of refugees and people smugglers, georeferencing census data and bathymetry. Precision agriculture seems to be talked about more than implemented although one case study was presented for a vineyard.
In the conference a number of different data sources were discussed. HERE, a company which provides mapping for in-car navigation and which collects vast quantities of data each day from the vehicles using its systems. The telephone companies, which have a huge consumer base, offer connectivity; persistent data with location; and they can map movement and behaviour in real time. This can be used to map population density, the flow of people and vehicles, location and routes of users; all of which can be used to carry out risk profiling and dwell time. Telco companies are becoming data companies, but success depends on partnership: Earth observation and geomatics are necessary components.
Other speakers discussed the use of imagery from drones and from terrestrial cameras. Stuart Martin, CEO of the UK Satellite Applications Catapult proposed learning from nature, saying that we need to adapt to environment, evolve, build healthy partnerships and become resilient in diversity. He noted the key trends in geospatial data as being feature extraction, classification, 3D reconstruction, change detection and forecasting. He suggested that a future use would be artificial intelligence to control satellite operations, for example to find tankers, and cited the use of many types of geospatial data to detect illegal fishing.
An inspirational presentation came from Jeff Jonas, a data scientist, and former IBM Fellow, entitled ‘Big Data. New Physics. And Geospatial Data Is Analytics Superfood’. He stressed the importance of context in order to interpret big data but defined the superfood as the ‘when/where’ provided by geospatial data, and this put together with telco data and satnav data could provide a very personal profile which would include routes and information on who you meet.
The message put over by most of the speakers was that there are vast amounts of geospatial data and that this is increasing, and that there is a demand for information for decision making, which could only be met by the use of artificial intelligence and collaboration through partnerships. Geomatics professionals are already familiar with this trend through BIM, but the additional ingredient which came from this meeting, is innovation.
Better DEMs for Climbers ad Skiers
Many of the presentations came from start-up companies, one of particular interest is FATMAP, a UK company set up to provide high-resolution DEMs of mountainous areas used by climbers and skiers. Finding that existing DEMs were not good enough and could not be displayed adequately on handheld devices, FATMAP used the basic DEM data and computer gaming engines to display it. In order to access the necessary new data a partnership was established with DigitalGlobe using Software as a Service (SaaS) through the cloud.
Here is an example of entrepreneurship taking very basic geospatial data and techniques but using software and business models from outside of the geospatial industry. Maybe the big opportunities for geomatics could come from looking at how geospatial data can be used with other data and software for new and innovative applications, these could include aspects of virtual reality, augmented reality and BIM. Operating autonomous vehicles requires very accurate geospatial data and artificial intelligence is a key component of their software, so there are significant opportunities in this area. Funding for innovative projects using Earth observation is available through the UK Satellite Applications Catapult.
This article was published in Geomatics World July/August 2017Last updated: 19/02/2018