Objectives of the service
The key project objective is to create and deliver “activity index” products tailored to customer requirements with a scale, frequency, accuracy, and price point never seen before. The service will give customers a much-desired competitive edge for trading in goods and commodities.
The activity indices will derive from a variety of types of activity indicators, for example:
Monitoring thermal emissions from sites including slag heaps and furnaces
Detection of smoke and exhaust plumes emanating from smokestacks at the site
Multi sensor change detection to detect and monitor freight transportation, including trains, containers and road vehicles
Detection and monitoring changes in stockpiles of input material
Detection and monitoring changes in holdings of produced material, where visible
In addition to these activity indicators, we will also explore the integration of 3rd party geospatial data to drive an alerts system to pre-empt activity outages caused by extreme weather or pollution.
Users and their needs
For physical commodity traders and asset owners, GAINS will provide activity data down to individual asset performance level. Customers in this segment often have very detailed information relating to assets they have direct commercial interaction with (or own). This allows them to benchmark GAINS data against their own inputs in order to interpret results for other sites they do not have information on.
For paper and systematic traders, GAINS will provide activity data aggregated up to a macro-economic level. This customer segment tends to be less interested in single asset data due to the difficulty in finding investment products with that specific exposure. However, datasets that can provide timely macro-economic insights for an industrial sector are of significant interest to this type of trader.
For market research firms, GAINS will provide an additional data point from which to inform and enrich their research. For market regulators, GAINS will provide a means of understanding market movements better, as well as for validating information being reported by the operators of industrial sites.
Service/ system concept
The service will focus on providing high-frequency activity indicators of complex supply chains by proactively monitoring and tracking the inferred levels of activity at diverse industrial sites – e.g. from thermal emissions, changes to assets and stockpiles, prevalence of night lights, evidence of lorries, trains and ships being loaded, atmospheric emissions and more. It will focus on industrial sites that produce tradeable goods and high-value commodities. The inferred activity levels will be correlated with complementary data sources such as market trading data and available published data on production from each site, as well as publicly accessible information such as news feeds, company press releases and trading statements. The specific information that will be made available to customers is a temporal graphical plot of the level of activity detected at each industrial site.
The cloud-based architecture will be a serverless model to provide scalability with reduced operational costs. The system shall import data of various types and format, which will trigger a predefined processing workflow. Each workflow is highly customisable and constructed from a library of functions. This provides the flexibility to create new workflows or modify existing ones. Depending on the workflow executed, the resulting outputs will be data stored in a database or files written to storage. A user interface will be developed that permits the search and retrieval of data using our custom API; advanced users could use the API directly within their own software.
Space Added Value
GAINS will use very-high, high, medium and low resolution optical, infrared and radar Earth Observation imagery and video datasets from free and open sources such as the Copernicus Sentinels, ALOS, VIIRS and Landsat; as well as commercial data from operators such as Airbus, Planet, Iceye, Capella, GHGsat and others. Our approach will be to start with the optical and infra-red satellites and datasets, then gradually expand into data types such as radar and thermal.
The programme will also be designed to accommodate future satellite and terrestrial remote sensing sources as they become operational, and additional open-source datasets as they become available. For example, new Copernicus Sentinels are likely to provide additional valuable datasets that will be complementary to the existing ones for this service. We will also investigate use of other space-derived information like weather data feeds, as severe weather events can also impact industrial output.
Since the project kick-off in October 2021, the Earth-i and Marex project teams have held a number of detailed requirements meetings with potential customers to gather their needs and expectations from the GAINS service, and details have been captured and documented within a Requirements Document.
Exploratory proof-of-concept development work has been undertaken on key capabilities that will be required to deliver the service. These proof-of-concepts focus on functions such as nightlights monitoring, stockpile monitoring, SAR change detection and stockpile change detection.