Geostatistics comprises a collection of statistical methods used to model spatial/temporal data and discern patterns in geographical phenomena, which by nature are subject to space-related and time-related variations. These techniques automate the correlation of sampled and unsampled data, enabling us to make generalizations about a dataset, e.g. processes, events and entities in a specific area.
Spatial/temporal processes and events are “moving targets”, i.e. they are incompletely known. Geostatistics makes extensive use of informed guesstimates and variables to analyze variations and deviations in these entities and phenomena, and extract information which would be otherwise difficult to pin down.
The science was originally developed for the mining industry to predict the distribution of ore grades, for instance, what is the possibility this area is rich in minerals? It uses random variable theory to model and overcome the uncertainties associated with spatial/temporal data.
While the measurements of the tiniest pieces of data in a sample are extremely accurate, over a large area or long duration, a sample cannot be measured in its entirety, which is where geostatistics comes in: to abstract from the measurements we have and come to an informed decision. The output from geostatical analysis are probabilistic models and patterns, in short: trends.
Interpolation is a process that uses measured samples to estimate probable values for unsampled locations. For instance, researchers use polls and surveys to study a sample of the population, and extrapolate from the findings certain trends. Geostatistics differs from traditional statistical analysis in that it tries to illustrate how reflective the results from samples are of a dataset, e.g. the entire population.
Geostatistics In Action
Geostatistics plots patterns, highlighting anomalies in data and relationships between entities that may be obscure using traditional analytic models. It focuses on statistics, and spatial patterns and relationships, and is used to create predictability, probability and error maps.
Business: At the heart of geostatistics analysis is geography. Outside of the scientific world, for businesses selling products and services, geolocation is an integral factor in analyzing consumer buying patterns and targeting specific demographic markets.
Meteorological: Can be used to measure a phenomenon like ozone depletion and plot trends in different geographical areas. These measurements can be used in conjunction with other location-based factors to extrapolate information about the causes of ozone depletion, e.g. weather patterns, regional elevation or pollution.
Weather Maps: Geostatistical analysis can create weather maps, using geographical factoring, to predict temperatures based on geographical factors, e.g. area forestation, proximity to the ocean or cities, location in the hemisphere, time of year and area elevation.
Disaster Management: Statistics enable scientists to predict the fallout, e.g. the probability of soil contamination, from disasters such as nuclear power plant accidents.
Crime: City law enforcement uses geostatistical methods to identify high crime areas.
AI: Spatio-temporal reasoning is used to “teach” robots how to navigate time and space, organize and manipulate data, identify patterns and visualize complex step-by-step processes.
In this tutorial you’ll discover how to build a geolocation tracking app that allows users to share locations and send messages in real-time.