August 25, 2017
Monitoring consumer behaviour with satellites
Tom owns his own analytics company that helps businesses drive and gain better insight from their sales and marketing data. The overall goal of his business is to draw on multiple data points to determine why things happen the way they do. His company ingests multiple data sources, such as satellite, social media sentiment, and weather, while allowing customers to add their own sales and marketing data as well. As a result, Tom’s customers have a much clearer picture of what is driving performance across different locations, from a macro level, such as in a city or region, down to a particular store.
Traditionally, satellite data has been difficult to work with for a variety of reasons. Buying satellite data can be very expensive as vendors require large minimum purchases, often resulting in customers paying for more than is required. Additionally, the variety of data formats and vendor technologies makes it extremely time consuming and costly to search and integrate satellite data from multiple satellite providers into a single application. Also limiting the amount of data sources used, to save on infrastructure costs, can mean there are just not enough data over a specific area or date range to gather the necessary information.
But finding an efficient way to access multiple satellite sources was not Tom’s only challenge.
CHALLENGE #1 - RECURRING DATA NEEDS
Tom also needed to access Earth observation data at a set frequency. He did not only need past data but also future data and did not wish to come back every week and have to recreate the same identical search.
EarthCache fully supports the set-up of subscriptions based around your own defined areas of interest, taking in account your minimum-required resolution, start and end date, preferred post-processing options, cloud cover threshold, and more . Whenever new data becomes available that matches your configuration, your team or application will be automatically notified, saving valuable data research time.
CHALLENGE #2 - BUILDING FOR MACHINES
For a long time, the absence of standardized format and unified points of access for remote sensing data made it nearly impossible to deliver anything other than stand-alone analysis from remote-sensing data. But this is changing fast. With the rise of cloud services and machine learning, humans’ interpretations of satellite data is becoming less crucial as derived features from data can be efficiently delivered at growing scale.
For his customers, Tom needed to be able to ingest Earth observation into his automated workflows quickly and efficiently.
EarthCache is designed to facilitate easy integration of high-resolution satellite data into any development environment in an Machine-to-Machine (M2M) environment.
Some of our new features include multiple input formats, such as KML and GeoJSON — manual selection is still available —, standard GeoTIFF output, additional free post-processing options, as well as a newly redesigned interface that will generate API calls for you to paste into your applications.
CHALLENGE #3 - A NEW DATA PURCHASE MODEL
In the past, purchasing satellite data was often reserved for governments or large corporations as the cost could be prohibitive, often due to having to buy an entire satellite data tile to obtain information regarding a small area.
At SkyWatch, we do not believe in minimum purchases. By setting a subscription to an area through EarthCache, you ensure you only receive and pay for the data you need, whether it’s 1km2 or 895km2, making satellite earth observation data an even more attractive source of intelligence for a growing number of mission critical business applications.
These cost-savings helped turn satellite data into an attractive source of insights for Tom and his customers.
Using the high resolution data acquired through EarthCache, Tom’s company can now enable their customers to create better sales forecasts by counting cars in retailer parking lots. Further, using archival data from EarthCache, Tom’s platform can count cars within parking lots and make correlations to marketing campaigns and product messaging by demographics within each area.