Assessing the maturity level of geospatial business intelligence

Search giant Google says the advent of Google Maps familiarised people with the concept of data points on a map, and progressively more and more BI customers are expecting fully integrated maps and spatial analytics with their tools of choice. Seeing data in the context of its location often exposes information previously hidden in the raw data.Google has already highlighted that, at the click of a button, the analytics currently being viewed by a user can be interactively viewed on Google Earth. All the information and data relationship layers are dynamically available so that as filters are applied in the BI environment they are automatically applied to the view in Google Earth.Assessing the contribution, Jean-Sebastien Turcotte, executive vice president/CTO, Korem, said that Google has helped in democratising mapping.  BI has traditionally excelled at who, when and what, but not the “where” component. But Geospatial Business Intelligence is more than just points on a map. While Google helps provide some context to point data, there are a lot of other pieces to this puzzle to deliver a strong analytical platform. In most cases, points are not sufficient as is business data. It needs to be augmented with third party data or analysis tools to provide a complete picture,” said Turcotte, who is scheduled to speak at the Enterprise Strategies for Location Intelligence USA 2011 conference to be held in Chicago on March 30th – 31st this year. 

 
Acknowledging the contribution made by Google Maps APIs mash-ups on the web, Luc Vaillancourt, CEO, Spatialytics, said Google Maps API provides effective and affordable mapping context, geocoding and routing capabilities, all the generic and common aspect of what location intelligence needs to provides other information systems.
 
“Pushpins on a map are not that different from dates on a calendar. It is very useful but it does not mean that you then master time, timing, tempo, rhythm, synchronisation, prioritisation and history. It’s the same with data on maps. BI is about analytics; geospatial business intelligence is supposed to provide spatial analytics,” Vaillancourt told TheWhereBusiness correspondent Ritesh Gupta. “It’s in the business logic and its data that the opportunity for innovation and growth lies for the industry players. Location data is flooding from mobility devices and social media but the OLAP nature of BI is about aggregation to allow drill-down navigation and discovery ... points (pushpins) are not appropriate for that and we see more and more options to handle lines and polygons with the Web-giant APIs.”
 
Progress
 
Overall, Turcotte said that while the technologies and data have been available for years, there have been no major developments specifically in the geospatial business intelligence arena. 
 
“What we are seeing, though, is a lot of new companies showing interest in geospatial business intelligence, especially in the form of geospatial dashboards. These tools are deployed through web-based interfaces and made available to a large internal audience,” said Turcotte.
 
“We (now) see more location intelligence in business intelligence,” said Vaillancourt.
 
Vaillancourt added: “What makes Spatialytics unique in the LI space is the depth of our integration, in all the data value-chain of a full BI system. Spatialytics integrates the geographic representation of information (geometry of facts and aggregations) at the core of the OLAP data warehouse and analysis server. Our Open Source Spatial ETL GeoKettle is capable of feeding in the proper way our SOLAP (Spatial OLAP) Server GeoMondrian. End-users can then interact with the multi-dimensional data via dashboards made of charts, graphs and maps making bi-directional MDX queries to GeoMondrian.” 
 
He added: “The same reasons we see geometry in a Spatial RDBMS like Oracle Spatial or PostGIS, apply here with Spatial OLAP. If maintaining one system and data warehouse, using the power of Spatial SQL and benefiting from the intrinsic integration of the geographic dimension with the other dimensions stored as alphanumeric data and its geographic location and representation bring advantages, serious BI project with a good use of geography in the data logic should consider a GeoBI deployment based on a SOLAP infrastructure like the one Spatialytics can provide.”
 
“Spatialytics is aware that putting in place a new system, full GeoBI (Spatial ETL, Spatial OLAP Server with Spatial MDX interaction), is not always possible and that the challenge is also to maximise existing information systems (OLAP or OLTP). In those cases, we still bring GEO to the end-user (Geo-dashboards and reporting)... knowing the customer may eventually consolidate their infrastructure,” Vaillancourt said. 
 
From Turcotte’s company’s perspective, geospatial business intelligence is not considered to be a software implementation project. “We try to provide a good advisory approach to our customers that help them take advantage of their data is a geospatial way. Implementing GBI is and involved process from data integration, to analytical requirements to finally deliver a solution. We tend to promote a strategic consulting phase at the beginning of a project to really help identify the target system in regards to a customer’s current situation.”
 
Requirements 
 
According to Turcotte, most companies want to implement geospatial business intelligence technologies, but don’t really understand the requirements. 
 
He referred to few aspects:
 
Is your data geo-reference (geocoded)? 
What level of accuracy do you need? 
What coverage do you need? 
Is there existing 3rd party data that could complement your data? 
Is the data structured appropriately for use in a GBI solution? 
 
“All of these factors will influence the price and complexity of data integration,” said Turcotte. “The best way to approach these issues is to start early in the GBI implementation process to roll in expert resources that are able to provide insight. Also, there are tools out there that help in cleaning and streamlining data, and the implementation of such tools will help in ensuring accurate analysis throughout the life cycle of the application.”
 
Decision-making
 
For his part, Vaillancourt says geospatial business intelligence is BI with geospatial intelligence. It insists on a dimension of the information that BI should include and handle de facto. It uses the same BI technologies but carries geospatial all the way, from the data integration, through OLAP analysis and data mining exploration, to end-users applications like dashboards and reporting. Beyond location from LI, geospatial business intelligence can play with OLAP, masters thematic mapping and allows spatial analysis, he said. It can generate map-centric or map-enabled BI systems according to the requirements from the vertical industry using it.
 
Commenting on how the industry has fulfilled demand from business users for better decision-making tools with fully integrated location and business intelligence functions, Vaillancourt said: “I think the demand in geospatial business intelligence, as we define it, is yet to come, mainly for complexity and cost reasons. This is why we are taking the dual approach of Open Source, via spatialytics.org, to let hard-core developers/integrators multiply instances and a commercial Open Source one, including support, addressing ease of deployment, improving the out-of-the-box experience and also figuring out a self-service GeoBI SaaS/PaaS offering.”
 
PaaS environment
 
The industry is placed to benefit from enterprise-wide business intelligence and geospatial analytics in a true Platform-as-a-Service (PaaS) environment. 
 
Turcotte said that while this idea may be sound in concept, there are barriers imposed by businesses. 
 
“BI or geospatial business intelligence tools are inherently tied to business operations, and the data that is represented in these tools is confidential and in constant flux. A cloud/SaaS approach to these types of project is often not even considered because of security issues raised by the companies. However, this mindset may change in the mid- to long-term future, but short-term I see these types of project being delivered internally,” he said. 
 
He added: “Also, I have yet to see a single implementation that is similar. So it is difficult to build a solution that can serve multiple customers. Maybe in a very specialised industry segments, but not in general. I think part of the reason is that geospatial business intelligence is fairly new and best practices and standards have yet to be adopted.”
 
Vaillancourt acknowledged that PaaS is a decent option to reduce cost, time and complexity associated with a geo-integration, but it depends on the depth of the integration. 

“Google Maps API have proved this but, most of the time, the integration was light (geography and advanced functionalities),” he said. “The more an organisation is involved with its own geography, needs to provide layers of geometry and the associated statistics and ask for advanced thematic mapping and spatial analysis, the less likely it will turn to PaaS, or it will be the last to join. It’s not only a technological challenge, it is a cultural one. The more an activity has in-house expertise and is core or near-core business, the less the company will turn to the Cloud.”

He added: “There will always be a place for an “on-premise” offering, but the “on-demand” one is growing fast and is facilitating the adoption of LI in general and will for geospatial business intelligence.”
 
Future
 
Turcotte sees the geospatial business intelligence arena as a growth sector. 
 
According to him, it will enable the integration of a geospatial mindset in new areas that have been historically less focused on geospatial technologies. 
 
“Of course existing geospatial users will keep requiring more and more advanced capabilities. Also, mobile deployment will be required for a lot of geospatial business intelligence implementations,” concluded Turcotte.
 
Jean-Sebastien Turcotte and Luc Vaillancourt are scheduled to speak at Enterprise Strategies in Location Intelligence USA 2011 to be held in Chicago this year (March 30-31). 
 Source :GISuser