Viewpoints
Why Teradata
Beyond numbers
Turn geospatial data into smarter decisions.
by Arlene Zaima and Ellen Boerger
To the uninitiated, it might seem like geospatial analysis would be useful only to the likes of NASA engineers or military commanders. However, much of the data stored in a corporate enterprise data warehouse (EDW) has a spatial component—including addresses, weather data, phone numbers and postal codes.
Until recently, there had been no easy way to integrate and analyze spatial and enterprise business data. Fortunately, technology has advanced to the point where companies can capture and analyze geospatial data in the EDW, adding a new aspect to business analytics.
Geospatial analysis
Geospatial data is precise location data represented by latitude and longitude coordinates. While data traditionally found in the EDW can reveal some location-related information, such as addresses or regional boundaries, that information has limited potential for tracking trends and making predictions. For example, most EDWs contain company or customer postal code data. However, a single postal reference can represent hundreds to thousands of different addresses. As these can span an enormous region, they fall far short of accurately pinpointing a location or mapping a specific area of influence.
Location data by itself is one-dimensional. Tying it to data about a company's products, services or customers is what enables in-depth analysis. With access to integrated geospatial and traditional EDW data, however, company analysts need not limit themselves to any pre-defined parameters. Geospatial analysis enables them to define virtually any area of interest they want to explore. In addition, company leaders can manage their organizations based on business-specific points, areas and locations.
Technological challenges
An explosion of geographic location data has occurred recently thanks to Web navigation sites like Google Earth and Global Positioning System (GPS) devices embedded in cars, cell phones and portable navigation units. Despite the prevalence of this type of data, many companies do not incorporate the information into their business analytics. This is because:
- Many different coordinate systems and file types exist for presenting geospatial data. In addition, thousands of geospatial reference systems calculate coordinates differently with algorithm variations among countries and geospatial product manufacturers. Manually attempting to reconcile these variations to produce business intelligence (BI) requires more resources than most companies are willing or able to expend.
- Not all data warehouses are capable of integrating geospatial and enterprise business data, much less dealing with the data transformation issues required to analyze it.
- The enormous amount of geospatial data generated typically exceeds what most systems can process in a timely manner.
Companies looking to add geospatial analytical capabilities to the data warehouse should make sure the solution they choose offers fast, in-database analysis of integrated traditional EDW and geospatial data. The product must also work seamlessly with leading third-party geospatial and data visualization tools and applications.
Industry examples
Businesses that want to glean the greatest strategic information from their company data will benefit from an EDW with geospatial analysis capabilities.
Communications
Network operations and communications engineering companies generate massive amounts of data every day. The ability to analyze that data quickly with a high level of granularity holds the key to reducing churn and increasing revenue through greater operational efficiency and improved quality of service.
Through geospatial analysis, companies can, for instance:
- Capture daily transactions from the network to identify areas that are experiencing a large number of failed connection attempts for voice, data, text or Internet
- Identify all complete and incomplete calls by time of day, location and subscriber to quickly diagnose problem areas and follow through with troubleshooting procedures
- Determine busy periods by location and call type so that capacity can be increased or hotspot areas rerouted to improve call throughput and service in areas of high call density
By integrating traditional and geospatial data in the EDW, companies of all types and sizes can add a valuable new dimension to their data analysis.
Insurance
Insurance companies rely heavily on predictive analysis and risk measurements. The ability to use geospatial data in addition to traditional EDW data offers this industry a rich vein of analytical knowledge that was previously unavailable.
Geospatial data can help insurance companies:
- Quickly perform density analysis to identify the number of policies written for the same block near a high-risk area so that the insurer can mitigate that risk
- Easily aggregate and analyze an area's geography, topography, construction, history of claims, and details of the risk zone for flood, earthquake or other issues to accurately price a policy
Retail banking
The retail banking industry is under enormous pressure to reduce costs and offer greater individualized tools and services to customers. Geospatial analysis can help bank leaders achieve these goals faster and with greater data transparency.
Among other possibilities, geospatial analysis enables this industry to:
- Perform a robust analysis of ATM use based on location, demographics and proximity to other businesses to help ensure adequate placement and optimal usage
- Conduct a mortgage density analysis to ensure the bank does not hold the mortgages for an entire block of retail stores; such a case could expose the bank to undue risk or losses, because when one anchor store leaves or fails, it often has a negative impact on others in the area
- Determine whether branch locations are optimized to attract and retain the highest-value customers—analysis could include comparison of branch and customer location demographics, services offered and peak period staffing requirements over time
Government
Along with their supporting agencies, governments generate vast amounts of data—exponentially more than the average enterprise. Because of the enormous size and complexity of this data, these entities have an even greater need to use geospatial data to help quickly perform the most accurate analysis possible. Like other large industries, government can benefit greatly from the geospatial visualization tools that make understanding and synthesizing large volumes of data easy and intuitive.
For example, integrated geospatial analysis can let government agencies:
- Solve or prevent transportation problems quickly by providing tools to predict or measure the impact that shifts in travel demand or capacity have on infrastructure resources
- Help military aircraft operators predict the time between repairs on key aircraft parts based on how they have been historically affected by weather conditions and geographical areas
- Aid healthcare workers in determining disease patterns and the efficacy of treatments across large populations over time by visually overlaying disease types and total number of cases within high-risk areas such as those with exposure to communicable diseases
Retail and hospitality
Company leaders in the retail and hospitality industry already know the value of analyzing who their customers are, as well as why and how they make certain purchases. Incorporating geospatial data into this analysis helps these organizations easily answer other questions that were previously too difficult or impossible to pin down, including, "Where are my customers driving from, and what products or services entice them to make long journeys?" and "Which loyal and highly profitable customers are 'at risk' due to a competitor's store opening in the area?"
Other uses of geospatial data for the retail and hospitality industry include the opportunity to:
- Accurately forecast the competitive position of stores over time to improve targeted marketing campaigns, set product price points and drive product assortment
- Optimize the prices and boundaries of delivery service areas to focus on the largest set of customers for the greatest profit
Added value
By integrating traditional and geospatial data in the EDW, companies of all types and sizes can add a valuable new dimension to their data analysis. This opens for them countless new possibilities to meet strategic initiatives, increase competitiveness and thrive with confidence.
Arlene Zaima is an advanced analytics program manager at Teradata and has been involved in data mining and related technology for 10 years.
Ellen Boerger is a global industry solutions director at Teradata with a focus on the retail industry.