Using Satellite Imagery + AI to predict financial health of an institution
Submitted by Gaurav Kaila
Signed on Ethereum on 9/23/2019
An AI application for predicting revenues of a brick and mortar institution by counting number of cars in its parking lot which uses a combination of satellite imagery, advanced AI methods and socioeconomic factors such as household income, unemployment, crime rate, etc in order to predict revenues before the quarterly/annual reports. Users benefit by having visibility of an institutions financial health in advance when making key investment decisions.
One of the key criteria for the success of any business is the number of active consumers it has. Social media companies such as Facebook release their active monthly users every quarter to showcase their stronghold in their market segment. Similarly, brick and mortar institutions such as shopping malls, grocery stores, amusement parks, etc have their success directly co-related to the number of customers that they attract.

There are a couple of ways to predict the number of customers an institution attracts.
1. By manually counting the number of people visiting
2. By card/cash transactions made daily/weekly/monthly
3. By counting people using CCTV footage

All the above methods do provide an estimate of the number of consumers but require manual effort that can be expensive.

My proposed idea is to use Sattelite imagery and state-of-the-art AI methods to count cars in an institutions parking lot and combine that with the socioeconomic factors such as the areas unemployment numbers, household income, crime rate, etc to predict the institutions' financial health.

This idea is not entirely new, as, since 2010, analysts have been looking at satellite imagery to aggregate financial numbers of an institution ( Though most of this has been manual counting due to the poor resolution of the images, now with the availability of high-resolution imagery, automation of this process is possible.

There are already similar ideas such as predicting crop yields using satellite imagery (,

There are other use-cases where such a technology could be used such as to predict a competitors financial health e.g. Burger King employing this tech to count the number of cars outside every McDonalds in a city.