Home Data Science Looking to the Skies for Alpha
Satellite Imagery
Highly detailed epic sunrise over world skyline. Planet earth Europe zone with night time city. 3D Rendering animation using satellite imagery (NASA)

Looking to the Skies for Alpha

0
2820
Thumb1

We recently discussed how companies such as Quandl are leading this alternative data revolution. Now, quantitative hedge funds are only too willing to spend substantial amounts of time, money and effort obtaining such data, all to gain that crucial informational advantage and generate alpha. And one of the most useful sources of alternative data is satellite imagery.

The number of satellites orbiting our planet is growing rapidly. Whether it’s for commercial use, or for government, military or civil purposes, we are seeing more objects being launched into space with every passing year. The Index of Objects Launched into Outer Space, as maintained by the United Nations Office for Outer Space Affairs (UNOOSA), counts 4,857 satellites currently orbiting the planet in 2018. That’s almost 5% more than in 2017, which itself was 9% more than in 2016.

With more satellites now available to map the length and breadth of our planet, therefore, hedge funds are increasingly using the imagery to identify significant economic change across key geographical areas, which can ultimately deliver potent trading signals. Indeed, the last few years have seen companies send massive fleets of satellites into orbit that can return streams of images of the Earth’s surface while having them continuously updated. Investing decisions are made on the back of images mapping the growth of crop yields, for example, or from identifying those mines producing the most coal.

Spy Satellite Digital Bird Eye
Spy Satellite Digital Bird Eye View – Search For Suspicious Car in Afternoon Commute. Digital Spy Targeting Theme. Surveillance Systems.

UBS was one of the first to employ this strategy. Back in 2010, it began using private satellite companies to collate the flow of Walmart parking lot usage. By observing the volume of cars coming in and out of the parking lots each month, the Swiss bank was able to project customer flow, and compile a regression analysis to forecast the hypermarket’s quarterly sales.

The finance industry’s demand for satellite imagery has only grown since then, with hedge funds increasingly using it to keep tabs on risks related to mines, ports, plantations and farmland. The power of the imagery itself has also improved dramatically, which is enabling more granular data to be identified. The quality of ore and crop output can also be further analyzed using infrared technology. As such, hedge funds are now only too keen to snap up this data before their peers.

And for good reason. Countries are not always willing to provide data, especially from the more sensitive or impactful spheres such as crude oil supply. Nor are they always accurate when presenting their official data – China springs to mind in that regard. And so, the pressing need to gain more honest assessments from satellite imagery becomes apparent. What can birds-eye images of the earth tell us about China’s economic activity? Or about oil storage levels? And how can it be translated into actionable forecasts of market movements?

That’s where a company such as Orbital Insight, which has been offering its clients trading signals based on its analysis of satellite imagery since 2014, can help. “Orbital Insight builds a macroscope and finding truth and transparency in the ever-expanding supply of satellite images.” In reality, this means the company employs data scientists and engineers to source, process, and analyze satellite imagery to make it actionable for a diverse client base that includes hedge funds, insurance companies, governments, and NGOs. By 2016, Orbital Insight was being used by 70 hedge funds and 4 government agencies.

geospatial

source

Orbital itself is involved more at the data analytics stage, rather than at the satellite imagery stage. As CEO James Crawford explained, “We take imagery from all satellite companies around the world, put it into one artificial intelligence-based system and do a tremendous amount of data science on top of it to finally come out with an understanding of what’s going on in the world.” But in terms of actually obtaining the imagery itself, Orbital has signed agreements with various satellite companies – among the most exciting is Planet Labs.

Founded by former NASA employees, Planet Labs’ mission since 2011 has been to image the entire Earth land area every day. By February 2017, having launched a record 88 shoebox-size “Dove” satellites, it had amassed the required number of satellites (149) to achieve this feat. Along with ‘cubesats,’ the small satellites can deliver more frequent images of target areas of the globe than traditional satellites – areas such as oil-storage facilities, parking lots and crop land. Today, Planet Labs processes around 2 petabytes (2 million GB) per year.

February 2017 also saw satellite imagine company SpaceKnow raise $4 million in a funding round led by German venture capital firm BlueYard, in order to develop AI tools that can “scan, understand and describe every square inch of the Earth surface daily using satellite imagery”. Today, the company is best known for having developed the following:

  1. The China Satellite Manufacturing Index, which combines satellite imagery with street maps and spectral data to monitor activity at 6,000 industrial sites in China. The index uses 2.2 billion satellite snapshots of the sites, covering 0.5 million km2 of territory to gauge the performance of the country’s manufacturing sector.
  2. Africa Night Light Index, which uses the intensity of illumination in individual nations and the continent as a whole as an economic indicator.

city receiver

 

The diversity of data that SpaceKnow is now extracting from the roughly 200 satellites to which it currently has access is startling. Machine-learning algorithms can analyze imagery of inventory goods, real estate, surface materials and much more, while co-founder Pavel Machalek recently discussed the example of images showing grass-covered land being gradually converted to concrete as a solid indicator of manufacturing expansion taking place.

Indeed, it’s the process of converting those mountains of satellite imagery into meaningful, alpha-generating nuggets that is of most value to hedge funds. Although upcoming companies such as Spire Global and BlackSky now offer raw satellite imagery to monitor crops, parking lots, shipping traffic et al., much of the analytics work is left to the hedge fund customer.

But companies such as Orrery.ai are now striving to go one step further by implementing machine learning algorithms to process data and produce meaningful business insights. Although the company will begin by purchasing data from the likes of Planet Labs, it aims to put its own satellites into orbit. The small satellites known as ‘chipsats’ will be equipped with radar instruments at all times of the day, irrespective of weather conditions.

Eventually, the company intends to apply all the data it gleans from satellite imagery into investments that will be made through an allied 2020 Fund. As such, the fund could utilize Orrery.ai’s findings to make money for its investors. As founder Dick Rocket recently acknowledged, “There’s a gap between the data analysis firms and the financial sector…That is our gap.”

Given the significant reduction in the cost of launching satellites in recent years, it seems only likely that the amount of activity taking place in our world will continue to be covered more comprehensively. And that means the amount of potentially useful data that hedge funds have at their disposal will only continue to rise.

Could there be too much imagery to make sense of at some point in the future? Possibly. But in the meantime, it would seem that the hedge fund arms race for alternative data has only begun to accelerate.

Previous articleTechnical Setups – Forex
Next articleCross-border Payment systems: SWIFT, RippleNet or BWW?
Justin Chan
Dr Justin Chan founded datadriveninvestor.com, a tech application platform designed to help fund managers, individual investors and data scientists uncover outstanding sources of alpha from data. Previously, he specialized in strategy development in a number of hedge funds and served as a senior quantitative strategist at GMO. An expert in microstructure, market liquidity, and behavioral finance, Dr. Chan holds a doctoral degree from UCLA and served as a finance professor at Singapore Management University. He can be reached at justinchan@datadriveninvestor.com.

LEAVE A REPLY

Please enter your comment!
Please enter your name here

HTML Snippets Powered By : XYZScripts.com