What is Spatial Analytics?

Spatial analytics deals with data relating to the position, size, or shape of items in (two or three-dimensional) spaces. It is often used synonymously with geospatial, geographical, or location analytics. On closer examination, however, these terms refer to features and phenomena on or near the Earth's surface and can be viewed as a subset of the broader term “spatial analytics” [1,2].

Figure 1.  Geospatial analytics is a subset of spatial analytics

Spatial analytics has a broad scope. As Figure 2 depicts, the foundation for all spatial analytics scenarios lays in the location of one or more physical objects. This basic feature can be useful, for instance, to visualize locations of items (e.g. a company’s fleet) at a certain point in time and calculate the distances between them. However, if this location information is stored and complemented with a time dimension, it becomes feasible to track items in time, which in turn enables many new applications, such as path analysis or the calculation of certain spatial-related KPIs. In a next step, advanced analytics and machine learning approaches can be used to analyze this data in order to reveal patterns, support decisions (e.g. optimization of routes), or even predict behavior.

Figure 2. Scope of spatial analytics use cases

To support this functionality, spatial analytics spans the entire business intelligence and analytics pipeline (see Figure 2):

1. The extraction, transformation and loading (ETL) of spatial data from various data sources like web services, open data, sensors, or from metadata in other analytics systems. This can either happen on a regular basis with batch processes or in (near) real-time with streaming technology.
2. The storage and management of spatial data in a spatial data warehouse (SDW), which serves as an integrated data repository, similar to a common data warehouse. However, SDWs are usually extended with specific structures and models optimized for storing and processing location data [3]. Moreover, they frequently encompass Big Data and NoSQL technologies to enable heavy-write or heavy-read scenarios [4].
3. The analysis and modelling of spatial data that hold various challenges such as large amounts of data, the need to combine spatial and non-spatial data sets, as well as the need to support complex geo algorithms. Hence, there is extensive research on the efficient implementation of spatial data cubes and the exploitation of upcoming technologies like graph databases.
4. The visualization and integration of spatial data that finally present the processed data in an adequate form to facilitate insights and support decision making. For this, analytics and dashboard tools usually provide widgets that, for instance, visualize data on a map. Moreover, there are numerous specific tools that allow more sophisticated visualizations or visual data mining on spatial data.

Figure 3. Layers of Spatial Analytics

How can the Internet of Things leverage Spatial Analytics?

Spatial analytics heavily depends on its available data and consequently becomes more valuable the more completely and more precisely physical locations can be captured. This is where the Internet of Things (IoT) comes into play. Firstly, positioning sensors are one of the most common sensors in the IoT. They allow to exactly locate and track objects by providing absolute (GPS) or relative displacement information. Secondly, the large number of smart objects in the IoT and their ability to send data via mobile networks allow getting comprehensive and up-to-date datasets. This is why the IoT is a major driver for spatial analytics scenarios.

The examples below illustrate the broad possibilities that the IoT provide for spatial analytics.

1. Crisis and Disaster Management

A key challenge in crisis and disaster management is to stay on top of things. Here, the IoT, spatial analytics and geo visualization techniques can support collaboration, problem solving, and decision making. Smart objects provide real-time data about the location of people (via smartphones or wearables), resources (via connected cars or smart tools) as well as information about surrounding parameters (via sensors in a smart city). The major challenge is to combine all these sources and visualize them in an adequate form in order to provide a holistic perspective on a situation and allow reasonable decisions. However, once such an infrastructure is in place, it can not only help in a crisis, like during the Haitian Earthquake in 2010 [5], but also help to prevent it, e.g. with a smart water monitoring system in disaster prone areas [6,7].

2. Smart Cities and Public Services

With the arise of smart cities and open data, there is more and more information about urban areas and their inhabitants. Most of the data is related to certain locations in a city and can thereby be used for spatial analytics. There are many promising applications like analyzing neighborhoods and infrastructure for urban development or optimizing public transportation by analyzing taxi rides [8], to just name a few.

Due to the great number of data sources and good mobile network coverage, there are also many use cases with a rather operational character in urban areas. A great example is predictive policing, where police departments and other emergency institutions analyze past and real-time data in order to identify hotspots and manage their physical resources [9].

3. Indoor Spatial Analytics

All the examples above have a strong geospatial character. However, spatial analytics goes beyond mapping data on a map. Indoor spatial analytics, for instance, tracks the location of items in a building. In contrast to the scenarios above, indoor spatial data is usually relative and does not originate from GPS sensors, but rather from specific sensors like RFID chips in shopping carts, triangulation of wifi or bluetooth signals, or video cameras.

A common use-case here is customers’ flow analyses in retail stores that are used to understand the behavior and paths of customers in stores in order to improve conversion. Of course, there are many other use cases, for instance sports analytics, where researchers mapped spatial shooting data from NBA games to a virtual court in order to analyze the shooting behaviors of players [11].

Conclusion

The article showed that spatial analytics in combination with the IoT can be more than just dots on a map and companies should be aware of their spatial data and its potential. Moreover, the use cases illustrated the broad range of applications that become feasible with IoT technology and that it is sometimes necessary to think outside the box.

However, the application of spatial analytics in the IoT also holds challenges. First, large amounts of data have to be extracted, stored, and processed. Second, there are many heterogeneous data sources to be integrated. Third, many operational scenarios require real-time data processing. And lastly, there is the need for reasonable visualizations.

In order to approach these challenges, companies, on the one hand, should extend their BIA infrastructures with adequate technologies, like cloud-based spatial data warehouses or streaming platforms. On the other hand, it is necessary to build up specific governance structures and skill sets to get the most out of their spatial data.

Further reading

[1] (2010) Bhatta, B.: “Geographic(al) Information System, GIScience, Geomatics, Geoinformatics, Geoinformation Technology and Geospatial Technology”
http://basudebbhatta.blogspot.de/2010/02/geographical-information-system.html 

[2] https://gis.stackexchange.com/questions/34733/spatial-data-geodata-geographic-data-geospatial-data

[3] (2000) Stefanovic, N. et. al.: “Object-based selective materialization for efficient implementation of spatial data cubes”

http://ieeexplore.ieee.org/document/895803/

[4] (2014) Pourabbas, E.: “Geographical Information Systems: Trends and Technologies”

[5] (2010) Zook, M. et. al.: “Volunteered Geographic Information and Crowdsourcing Disaster Relief: A Case Study of the Haitian Earthquake”

http://onlinelibrary.wiley.com/doi/10.2202/1948-4682.1069/abstract

[6] (2017) Dunaway, M. et. al.: "Research Agenda in Intelligent Infrastructure to Enhance Disaster Management, Community Resilience and Public Safety”

https://arxiv.org/abs/1705.01985

[7] (2015) Perumal, T. et. al.: “Internet of Things (IoT) enabled water monitoring system”

http://ieeexplore.ieee.org/abstract/document/7398710/

[8] (2015) Richly, K. et. al.: “Optimizing Routes of Public Transportation Systems by Analyzing the Data of Taxi Rides”

https://drive.google.com/file/d/0B3IF1hOFyAo0M21MbU9lUUtzcEU/view

[9] (2013) Homeland Security: “Geographic Information Systems and Predictive Policing Application Note”

https://www.dhs.gov/sites/default/files/publications/GIS-Predictive-Policing-AppN_0813-508_0.pdf

[10] (2015) Wu, Y-k. et. al.: “Customer's Flow Analysis in Physical Retail Store”

http://www.sciencedirect.com/science/article/pii/S2351978915006733

[11] (2012) Goldsberry, K.: “CourtVision: New Visual and Spatial Analytics for the NBA”

http://www.sloansportsconference.com/wp-content/uploads/2012/02/Goldsberry_Sloan_Submission.pdf