Alternative Data Category Description
Satellite data, also referred to geospatial data, is alternative data that includes photographic images collected by satellites orbiting the earth and can be enhanced with data from drones and radar. It is a common category of alternative data used by asset managers as the interpretation of satellite images into data or intelligence is useful to asset managers on many fronts e.g., general economic activity, agriculture, mining, construction and real estate, shipping, oil and gas, project monitoring and retail.
Temperature, degree days, precipitation, visibility, wind, and solar radiation are a few meteorological factors that have the potential to impact energy, agriculture, and other weather-sensitive instruments. Weather data, obtained from meteorological satellites, can be utilized to understand these forecast variables, mitigate weather & climate risks, and uncover investment opportunities.
Asset managers first tried using satellite data back in 2010 with CNBC reporting that “Cold War-style satellite surveillance is being used to gather market-moving information”. According to the article, UBS Investment Research used data to count cars in Walmart parking lots to help estimate revenues, a strategy also implemented by Walmart’s founder Sam Walton using airplanes in the 1970s.
Alternative data providers that specialize in analyzing satellite imagery often use artificial intelligence techniques (e.g., deep learning) to come up with innovative ways to analyze satellite images. Asset managers use that information to assist in their estimation of revenues of companies in different sectors.
Satellite data is also seeing utilization in the data category of ESG. Satellites specifically designed to analyze emissions data can provide methane and CO2 emissions analysis. This data can go down in granularity to offshore emissions, measuring methane from cow burps, gas pipeline leaks, etc.
Back in 2010-15, satellite imagery data still need highly trained imagery analysts with specialized software experience to derive insights using statistical methods manually. Making this type of work was difficult to automate and scale. Recent years have seen a rise in the popularity of this data among asset managers as costs have declined. It has become cheaper to put satellites in orbit. Costs have also come down in processing and analytics due to the advances seen in machine learning techniques. With increasing numbers of commercial satellites in orbit and significantly lower costs to analyze the images, satellite data could become a mainstream data source.
Subcategory - Satellite Data
Datasets under this alternative data subcategory generally use images sourced from satellites that are mostly used for resource surveying, pollution monitoring purposes, monitoring natural catastrophes, and measuring economic output. Car count data is a classic and still very popular dataset for the financial vertical for predicting activity in particular areas, gauging the interest/footfall of retail locations, etc.
Generally speaking, satellite imagery is combined with other data sources to deliver actionable data insights. Data vendors could be using satellite imagery data combined with news and social media sources to provide global facility and activity monitoring service; combined with aircraft sensors data for greenhouse gas emission data; combined with AIS data to gauge traffic at marine ports, and many more different applications.
Technological advancements coupled with increased coverage of the planet will result in more frequent observations and more data points for models to work with. This will improve prediction accuracy for models which use earth observation data as an input variable, and potentially see more entrants into the market with niche datasets focusing on specific sectors and applications.
Subcategory - Weather Data
Weather data is separated out as an alternative data subcategory since this data is usually generated from meteorological satellites. The weather has a substantial impact on businesses. Weather and climate data can be utilized to understand the impact of weather on investable assets, manage and mitigate weather & climate risk, generate trading signals, portfolio alignment with climate trajectories, and scenario planning.
Funds with positions in soft commodities can use weather data to monitor evolving temperature and precipitation anomaly trends as well as forecasted anomaly trends to quantify the impact on commodity futures. Environmentally conscious funds can evaluate the potential for climate impacts in certain industries. Investment firms with a position in retail can use weather data to better understand how weather influences footfall traffic year over year at particular stores.
Data Structure
- Since this type of alternative data can be hard to cleanse and structure. Some companies that provide raw imagery data typically would not have their data mapped to tickers. Other datasets that track macroeconomic/commodity activities would by nature of data not have ticker mapping. Typically, satellite imagery data focused on retail, real estate and such would have their data mapped.
- History can vary. Alternative data providers that use open-source satellite imagery can have 10 or more years of history. Others buying satellite imagery from commercial satellites, or datasets with a more niche focus would have less history.
- Delivery frequency can range from daily to monthly or even quarterly. This takes account of the time for the data to be transmitted, collected, and processed.
Compliance Considerations
Material Non Public information and Personally Identifiable Information are not a concern for this type of alternative data. The one thing to focus on for this type of alternative data, like all alternative data, is data provenance. A data vendor may source data from multiple data sources and compliance, data provenance, and contractual relationships need to be examined closely.