The Five Most Genius Applications of Spatial Analytics for Businesses.
For all the business decisions that have the “where” question.
Location Intelligence is being used in innovative ways to change the whole game! This is the second post under the same theme How Location Intelligence Powers the Most Complex Decisions You can read the first part here.
For years, intuition guided most decisions — from where a chain’s next supermarket should be located to estimating stock market prices. Today, you won’t be able to survive with just intuition. Fortunately, Data Science, Machine Learning, and Geographic Information Systems (GIS) can power you to make the ‘where’ decisions you need to make.
How data is being leveraged today is fairly common knowledge. Since data collection happens in time and space, often the location component of this data can be just as valuable. However, as it requires additional tools and technology — Location Intelligence is a relatively untapped powerhouse. Companies in 2019 cannot afford to be intimidated by this minor barrier because the benefits of leveraging your data’s location component are enormous!
Let’s have a look at some ways Location Intelligence is being used to change the business landscape as we know it — some of these might seem obvious, but others will definitely surprise you!
1. Facebook: Bringing Internet Connectivity to everyone
In 2013, Facebook announced that it wanted to provide internet access to everyone in all corners of the globe. Connectivity can be provided by different approaches like by beaming from balloons or drones, using landlines or satellite signals — in different cases, you would need different methods to provide coverage, depending on the geography and population density of a region.
In 2016, the social media giant revealed that it created high-resolution population maps for 22 nations. In April 2019, it revealed that the project now covers “the majority of the African continent”.
High-resolution images of the regions can easily be provided by satellite data. But the images are meaningless if they cannot give us a distribution of population densities. The traditional way to do this is to get people to label buildings on the map and crossreference this with census data. Given the landmass the project is covering, it would be a colossal use of manpower and take a lot of time.
This is where the magic of Machine Learning on Satellite Imagery helps! Using machine learning models on population data sets and satellite imagery, Facebook engineers mapped millions of structures distributed across the landscape and using that extrapolated the local population. What would have taken them years and years of human hours took a fraction of the effort!
2. UBS: Predicting Wal-Mart’s retail sales
What if you were an investor that could predict retail sales of a chain like Wal-Mart? Sounds crazy right? This is the exact superpower Neil Currie, an analyst at investment bank UBS, demonstrated in 2010 by predicting the earnings of Wal-Mart’s next quarter.
Using satellite data from more than 100 Wal-Mart stores and by counting the number of vehicles in their parking lots each month, he was able to understand the Wal-Mart customer flow.
He spotted a correlation with the monthly parking lot data and quarterly earnings for that month and so worked up a mathematical regression, a common math technique used by data scientists, to predict quarterly sales for the retail company.
What seemed like the work of a psychic at first turned out to be nothing but clever use of data science! The only real limit to how you can use data science and location intelligence is your own ingenuity!
3. Amazon: Optimising the business of physical stores
The online nature of Amazon allows it to escape a big problem that conventional stores faced — matching demand and supply of various items stocked. Have too many pink umbrellas and lose valuable storage space for green raincoats — which happened to trend more in that location during that month. Have too few green raincoats and lose out on the potential sales from the boosted demand.
But Amazon is also moving to the physical retail space because it realizes that not everything can be sold online (e.g. fashion) and that the future of retail is a blend of physical and online. Dominating the online retail space for years, Amazon has a steady source of consumer data with valuable location components which allows it to see these trends in real-time! And this is key to it dominating the physical market too!
Amazon overlaid geographic data on authors, sellers, and developers on a map of the United States. Around the same time, it had also launched its first physical bookstore. It leveraged this analyzed data to stock books according to what the people living nearby were interested in.
Today, the retail giant has many physical outlets in the form of Amazon Go stores. If a company like Amazon is going hyperlocal, other companies have little reason not to! The map below shows Amazon’s economic impact across each of the United States.
4. Predicting Corn Yields using Satellite Images
Knowing how much the crop yield will be is important information people in the agriculture industry would like to know in advance. It would provide answers to questions like when to sell their products and for how much — this can help those involved in agricultural distribution make better preparations and optimizations and ultimately make their business more profitable.
Joe Phongpreecha and team harnessed the power of data science to solve this problem! They knew that satellite images could be used to predict the type of crop growing in different regions — what was needed to do was predict the yield in those regions.
They collected satellite data of a county from different satellites from different years along with the yield in each year. Using data from 2010 to 2015 for training and data from 2016 for testing, they managed to find correlations between the images and the yield of each year.
They leveraged these correlations to construct a model that when given spatial and temporal components of data would give the predicted yield as output. If you’re interested in the details, you should definitely check out Joe’s article!
5. Revitalizing the petroleum industry
The petroleum business is an industry that is also the backbone of many other industries like the automobile, plastic and textile industries. Over the years, this age-old industry has adapted newer technologies into its arsenal and GIS has been one such technology.
Data index maps are one of the first things that come to mind when you think of GIS in petroleum. These maps allow one to easily see what is available and where and therefore they reduce the amount of time spent in searching.
The oil sector also deals with the extraction of other resources like shale gas and coal bed methane. Spatial analytics is being used to optimize for the most efficient drilling mechanism for a particular resource.
To ensure the punctual delivery of goods and services, there needs to be effective and precise tracking of the vehicles and vessels. Knowing vessel locations can also help in case of emergencies. GIS makes all this much easier.
Vehicles aren’t the only means of transportation in this sector — pipelines prove to be effective too. However, building pipelines can be an expensive task — therefore optimizations need to be made wherever possible. Fortunately, least-cost path analysis is relatively simple with GIS information. This analysis calculates the path of least cost by considering parameters like path length, the unevenness in the terrain, the cost of laying the pipes, etc. The analysis also results in more environment-friendly routes.
Location data has always been important and in the coming years, its value will only increase further because more and more location data is being collected. In 2020, there will be 20.4 billion connected ‘things’ as part of the Internet of Things revolution. Each of these things has a location and therefore the need for location intelligence will explode!
Do want to want to start using your location data in clever and powerful ways? Even GIS will be not be able to predict the upper bound of your potential!
If you want to delve further, check our website out or get in touch with me on LinkedIn or Twitter. Originally posted here.
Similar Reads
Written & Edited by Shawn Pereira!
