Business Intelligence Network - The Vision for BI and Beyond

“Location intelligence uses geographic information system (GIS) tools and techniques to transform and analyze data that becomes valuable information to make more informed and rational business decisions. Thus, location intelligence is neither simply an analytic technique nor business process; but a discipline that encompasses both.” – Encyclopedia of GIS

As the adoption of new technologies increases, there comes a time when features and capabilities transition from the area of the specialist into that of ubiquitous usage. Specialized products become features in broader applications:

• As the general market recognizes their value,
  
  
• As it becomes simpler to include them in other products, and
  
  
• As the costs associated with the technologies drop.

In past decades, we have seen data management layers in commercial applications replaced by database management systems; specialized word processors first subsumed by PCs, then largely incorporated into every application that uses text; and charts, spreadsheets, and sophisticated visualization and analysis tools incorporated in business intelligence (BI) suites and applications.

Today, there are two predominant approaches that are used to bring maps and proximity analysis into business intelligence and business applications. The first and most traditional approach uses integration with a geographic information system, or GIS. For decades, the GIS has been the exclusive domain of specialists, who developed stand-alone GIS applications or applications based on and integrated through the GIS platform. GIS specialists have traditionally viewed the GIS as the proper repository for location and geographic data, and typically associate essential business or operational data with the geographic information system as “attribute data.”

As a result, a class of applications and tools, parallel to standard business and BI applications, has emerged. While these applications are highly dependent on geographic information like locations, addresses, territories, census blocks, postal codes and coverage areas, they routinely rely on other business or operational information. This information – customer details, descriptions and history of assets, demographic data – is already found in other business applications and databases.

GIS-centric applications can support sophisticated detailed relationships and analysis using information relevant to most BI and business applications. But this approach depends upon redundant data, replicating data that is already available in other applications. Although the only data unique to the GIS is the detailed location content, the application must use specialized GIS software to access the data. Access to GIS data is often more difficult and may entail additional software licensing costs due to restrictive or proprietary access methods and interfaces.

The second commonly used method to integrate maps into BI applications is to access web-based mapping systems and display results through “mash-ups” with commodity, commercial or public sector location services (e.g. Google Maps, U.S. Geological Survey, Environmental Protection Agency, Traffic.com). The key advantages of BI location mash-ups are that they are relatively easy to integrate from a programming perspective. They are accessible through standard APIs that deliver high quality background mapping services. The limitations to this approach are that:

• Geographic access and analysis is limited to what is accessible through the APIs;
  
  
• Analysis limited to geographic hierarchies available on online service;
  
  
• It may be difficult or impossible to evaluate or even display the BI or application data you want to use;
  
  
• Your BI or application data may need to move from your system to an external system and back to perform geographic and relationship analysis; and
  
  
• Map design, look and feel are limited to that delivered by the map service.

A New Approach

With current web technologies, standards-based components simplify the incorporation of GIS-like capabilities into BI dashboards, reports and analytics. For example, Java Server Faces components now allow maps to be presented as chart types and user interface (UI) elements, provide thematic map components, enable geocoding (the process of transforming an address into a position on a map), proximity queries (how close or far is something) and containment queries (is something within a territory or region, and how many or what is the value of the items).

In this alternative approach to integrating maps and location analysis into BI and business applications, the web components invoke the geospatial capabilities (geocoding, proximity analysis, containment) found in modern database management systems that also contain most of the business and operational data needed by the BI application. This allows the BI system to generate any kind of query that might be required, no matter how detailed. In addition, map content may reside locally (in the same databases as other BI data) or may be accessed through a web service. The map component also offers the programmer complete control over the UI and map style, design, look and feel. This gives the application developer the freedom to integrate maps and location analysis using current web technologies, as an integral part of the BI solution.

In the pre-AJAX era, online mapping applications typically worked in a serialized single request/response model. More specifically, the browser/client issues a request for a map, and the server generates a map in response. The map is then displayed to the client. Any interaction on the client side, such as zooming or panning the map, will often result in a new request being sent to the server to produce a new map in response to the client action. In this previous paradigm, the server is constantly generating, and re-generating, maps on the fly. This often severely affects the scalability and performance of the overall application.

New web component-based mapping enables the caching of pre-generated map tiles on the server and in the client (via browser cache). New technologies provide map tile caching and many other unique features:

• Caching of map tiles for any base map at the mid-tier;
  
  
• Support for JavaScript (geospatial) objects generated at the mid-tier based on dynamic database queries and displayed at the client side with built-in interactions such as mouse over and tool tips; and
  
  
• Built-in capability to drag the map around, instantly zoom or pan to any desired level/location, draw freehand shapes on the map, add a collapsible birds-eye-view map inside the main map area, among many other off-the-shelf functions, available through open JavaScript APIs.

This approach offers powerful interactive mapping while hiding the complexity of spatial data queries and the cartographic rendering process from application developers. Users and developers can customize the appearance and behavior of the map; they can control visual map characteristics – such as the background color, the title, the symbology used to portray features such as roads, store locations and property boundaries, and so on using extensible metadata stored in database tables. It is also possible to incorporate dynamically obtained geospatial data, such as customer locations, and plot these on top of a base map. Thematic mapping portraying the distribution of attributes, such as population density, demographic information measuring income, education, etc., can also be supported through JavaScript APIs, if the user has the needed baseline data.

In conclusion, with web map components, every BI product can incorporate map visualization, proximity analysis, location-directed search, and real-time web services into the same infrastructure, and with no greater complexity than what is required for other visualization and analytic tools. Another factor influencing the cost and effort of implementing location-aware business intelligence applications is the expense associated with managing the data itself. Organizations are becoming aware that enterprise databases now enable them to manage and share geospatial data in the same database environment as their business and operational information.

Coupled with this web-based approach, consolidating the data in the enterprise database can further simplify the integration of maps and spatial analysis into business intelligence and other applications, while reducing or eliminating dependencies on specialty GIS software.

Using this approach allows geospatial analysis to transition from the realm of the specialty service into the area of a ubiquitous component that is dependent upon the data available for analysis, rather than a complex, single-purpose application.