Contrary to popular belief, geospatial data is all around us. So much so that it’s difficult to think of data that is not geospatial.
Geospatial data arises whenever events occur at specific geographic locations, such as disease cases in a specific area. Think of geospatial data as describing the what and where of events.
Space-time or spatio-temporal data describes the what, where, when an event occurred and is a more complete description of why the event happened. It provides a detailed understanding of what has transpired, the ability to forecast what will happen next, and insights into the process that underlies the data.
Let’s look at six examples where geospatial data is essential.
#1: Public Health
Space-time data is ubiquitous in public health and includes cases of cancer, the spread of infectious diseases such as influenza and monkeypox, and a host of others.
One example is late-stage diagnosis of breast cancer in Michigan. Late-stage diagnosis occurs when breast cancer is advanced before it’s diagnosed. These cancers are difficult to treat and result in a higher morbidity and mortality rate. Understanding where and when late-stage diagnoses occur allows for experts to establish more breast cancer screening facilities and communication programs that target local at-risk populations.
In collaboration with the Michigan Department of Community Health, BioMedware conducted several analyses (Goovaerts and Goovaerts, 2015) that revealed where late-stage breast cancers occur in Michigan. Overall the proportion of late-stage diagnoses declined significantly until 1999 when it started rising again; in particular late-stage diagnosis has been more prevalent over the years in the Thumb area (highlighted in the map below) where both access to screening (driving distance to mammography clinics) and socio-economic status are less favorable.

View of Vesta software where multiple visualization tools were used to explore how the percentage of women diagnosed late with breast cancer varies across the 68 counties of Lower Michigan and fluctuates between 1990 and 2006.
#2: Agriculture and Land Use Planning
Geospatial data can be used by agricultural planners to solve a number of problems in the industry including environmental health challenges, economic impacts of policy, and more. Let’s look at a specific crop where geospatial data is essential: cranberries.
Native to North America, cranberries are a high-value, intensively managed perennial crop that grows on wetlands. Strict federal guidelines prohibit the expansion of cranberry acreage on wetlands meaning that increasing profitability is most likely by more intensive or precise management.
Precision agriculture examines the spatial variation in crop yield, soil properties, and external factors such as weeds and disease and manages the variability within fields to maximize profit and minimize pollution. Minimizing pollution from fertilizers and pesticides is a particular concern in wetland areas whilst maintaining production levels.
BioMedware’s software and geostatistics helped data scientists:
1) Gain a better understanding of spatial and temporal patterns in usable and poor quality cranberry yield, leading to more precise management of the crop.
2) Identify deviations in crop yields recorded for different field owners and soil types—particularly in hot or wet years—illustrating the different responses of soil types to weather and the potential for improvement in irrigation practices by some owners (Kerry et al., 2017).
In this example, geospatial data was critical to identify “problem fields” with consistently low yields or within-field variability that would benefit from more precise management.

Screenshot of Vesta software where the standardized cranberry yield for the 1997 growing season was estimated by geostatistics for a few fields in New Jersey.
#3: Environmental disasters and risk assessment
Geospatial data is necessary for understanding the depth of environmental disasters. GIS software can produce in-depth digital maps that help emergency management professionals make quick and efficient decisions for the affected area.
Take the 2010 Deepwater Horizon (DWH) oil spill in the Gulf of Mexico, for example, which released roughly 4.9 million barrels of crude oil, and was by far the largest oil spill in US history. The legal settlement required an accurate assessment of the extent of contamination using data collected at various times and with different spatial resolutions and degrees of reliability, such as field measurements.
Stranded oil covering soil and plant stems in fragile Louisiana marshes was one of the most visible impacts of the DWH oil spill. Geospatial tools were essential to explore and model relationships between stem oiling field data and secondary information (oiling exposure category) collected during shoreline surveys as well as to estimate the death of vegetation, accelerated erosion, and other metrics of injury. Field measurements and secondary data were then combined using geostatistics to predict the probability of stem oiling at 50-meter intervals along the Louisiana shoreline (Goovaerts et al., 2016).

Screenshot of Vesta software with some of the data used to model vegetation history along the Louisiana shoreline: shoreline survey data on oiling exposure (left map) and field data on presence/ absence of oiled vegetation (right map).
#4: Retail
Retail store networks must be actively managed to know what’s in stock, predict sales, and prevent losses. Having the right goods in the right stores is critical to retail success and geospatial data can help ensure that happens.
Understanding which stores are excelling in sales, their location, and when this is happening drives stocking decisions, marketing programs, and performance assessments of individual stores across the network. This is critical in order to maximize return on investment for the retail network.
BioMedware applied its visualization and analysis capabilities to a complex space-time data series comprised of daily beer and wine sales in Dominick’s stores in the greater Chicago area in 1990. Statistical brushing was employed to identify an extraordinarily high volume of beer sales in an outlet in the central downtown area that upon further investigation turned out to be a data entry error. The user is brush selecting on the time series plot (top) to identify the spike in sales that occurred at one store in central Chicago (map, lower left) on Sept 23, 1990. Notice the strong periodicity caused by increased beer sales on weekends.

#5: Engineering and Infrastructure
New federal regulations require public water systems to have an inventory of all lead and lead-status-unknown service lines (SLs) by October 2024. However, most water systems tend to have heterogeneous data of uneven quality that are found in a mix of paper and electronic records of various formats.
Statistical modeling is a relatively new approach that supplements traditional SLs identification techniques, such as records screening, basic visual examination of indoor plumbing, water sampling, and excavation.
At BioMedware, we’ve developed, implemented and published the first geospatial approach to predict the probability that a tax parcel has lead or galvanized SL based on neighboring house inspection data on SL material and secondary information (e.g., built year and city records on SL composition).
Probability maps can be used to rank parcels according to their likelihood of having lead or galvanized service lines so you can prioritize inspection and line replacement. In the maps below, our model was validated using a dataset of 26,750 SLs that were excavated and inspected over a 5-year period (2016-2020) and represent close to 50% of all tax parcels in the city of Flint.

View of Vesta software where the probability of encountering a lead service line (left map) or a galvanized service line (right map) estimated by geostatistics for a few tax parcels in Flint, MI.
#6: Weather and climate change
GIS mapping and analysis is useful in identifying environmental changes so experts can implement a plan. In many developing countries, such as India, there is a great need for accurate estimates of rainfall, which is challenging as data collected at monitoring stations are either missing or costly, resulting in only spatial averages being available over large regions.
Whenever rainfall data is scarce, incorporating additional secondary information that’s correlated with rainfall and easy to acquire can improve spatial prediction. Satellite data, which has a high spatial resolution and is often downloadable at no cost, can be an attractive source of information.
With the help of Indian colleagues and BioMedware’s geostatistical tools, Dr. Goovaerts analyzed district-level precipitation data within the Indian State of Tamil Nadu. The data represent the average over 1901–2002 for the northeast (October–December) monsoon, which provides the major share of rainfall for Tamil Nadu. Potential evapotranspiration with a 1 km resolution was used as secondary information to map the spatial distribution of rainfall within each district (see right map below). Such a map of rainfall monsoon, which delivers about 70 percent of India’s annual rainfall, is critical for many sectors in India, starting with agriculture.

Screenshot of Vesta software where rainfall recorded during the northeast (October–December) monsoon for 29 districts of Tamil Nadu (left map) was downscaled to a finer resolution (right map) using geostatistics.
Geospatial data is everywhere, all of the time.
It’s clear that geospatial data is essential to all industries. Space-time data visualization and exploratory analysis further increases our understanding of the highly complex systems that make up the world we live in.
Try it for yourself with our software Vesta, visualization and exploratory space-time analysis software, and SpaceStat! Download both products here.
Sources:
Goovaerts, P., and M. Goovaerts 2015. Space-time analysis of late-stage breast cancer incidence in
Michigan. In P. Kanaroglou et al., editors, Spatial Analysis in Health Geography. Chapter 10, pages 161-180. Ashgate.
Kerry, R., Goovaerts, P., Gimenez, D. and P. Oudemans. 2017. Investigating temporal and spatial patterns of cranberry yield in New Jersey fields. Precision Agriculture, 18(4), 507-524.
Goovaerts, P., Wobus, C., Jones, R., and M. Rissing. 2016. Geospatial estimation of the impact of Deepwater Horizon Oil Spill on plant oiling along the Louisiana shorelines. Journal of Environmental Management, 180(15): 264-271.
Jacquez, GM. 2010. Space-Time Intelligence System Software for the Analysis of Complex Systems. In Handbook of Applied Spatial Analysis, DOI: 10.1007/978-3-642-03647-7_7.