Putting the 'where' into your data analytics
Success is a matter of the quality and quantity of data fed into the analytics system
By Sandra Gittlen | Computerworld US | Published: 19:00, 12 June 2012
At EDENS, a developer, owner and operator of community shopping centers on the East Coast, blending geographic information systems (GIS) and business analytics has enabled a competitive advantage in a fast-paced, crowded market.
The company coupled Microsoft SQL Server with Esri's ArcGIS Server and Spatial Database Engine (SDE) to display and analyse its current portfolio of more than a hundred properties and 4,000 other relevant shopping center property locations. EDENS' GIS director, David Beitz, integrates demographic data such as average household income, competition and traffic counts with select map criteria to paint a clear picture for both internal teams and leasing agents to use with potential retailers.
For example, grocery store chains have to closely study trade areas for new development. "They have found that some people just won't cross certain railroad tracks, impacting the site's overall draw," he says. With GIS-enhanced analytics, the development team can quickly rule out some sites -- even if they otherwise appear ideal -- that don't meet specific criteria. "We are able to really hone our presentations and give customers one or two great options. They can tell we've done our homework," he says.
Related Articles on Techworld
While it's difficult to quantify the impact of GIS data, success is measured by being able to quickly respond to retailers' requests for location and market data to help leasing agents "move a deal forward," Beitz explains. "If you make a retailer wait or provide out-of-date information, then they are more likely to land at another site."
EDENS finds that success is a matter of the quality and quantity of data fed into the analytics system. For instance, though relatively few data points are shared with customers, all information, including confidential prospectus data, is mapped and stored for later use. If the team wants to buy a property down the street from one it looked at in the past, all relevant data is at the ready, avoiding the need to reassemble all the past shopping center and market data. Rapid access makes determination of the soundness of the deal faster and more efficient, Beitz says.
In Las Vegas, engineers at VTN Consulting, a civil engineering and land planning firm, are beginning to use GIS and analytics to bolster communication and collaboration among project stakeholders. Their pilot project, based on Autodesk's AutoCAD Map 3D and Autodesk Infrastructure Modeler software, interweaves civil, geospatial and building data. Planners, GIS analysts, project managers, architects, city leaders and other stakeholders can visualize projects in context and with existing parameters, such as underground utility lines, traffic patterns and surrounding buildings.
"Traditionally, engineers have to work with multiple two-dimensional paper plans that are tucked away in different silos," says Keith Warren, VTN's Building Information Manager.
VTN's project comprises a central database that serves as a storehouse for survey and architecture plans, infrastructure specs and more. Pinpoints include water and sewer pipes, utilities, street signs, parcels, roadways and structures. Information is rendered in 3-D so that users can visually analyze the impact of new construction or renovations based on site requirements.
The striking models can help planners decide the number and location of street lamps to place outside a new building based on a light and shadow assessment of the structure and its environment. Or they can clearly illustrate utility thresholds for various proposed projects -- for instance, a library would probably consume less energy than a casino.
Warren says the database was built using standard GIS fields so that clients will be able to hook into their own analytics engines. For instance, if a building's air conditioner goes out, repairmen would not only be able to call up its exact location but whether nearby air conditioners also are in need of service based on their maintenance records.
Saving animals' lives
For the American Society for the Prevention of Cruelty to Animals (ASPCA), stepping up business analytics to include GIS data is literally a matter of life and death.
Millions of cats and dogs enter animal shelters annually, and more than half are euthanized due to non-health-related issues such as overcrowding. The ASPCA battles this overpopulation crisis with low-cost or free services, including spay/neuter clinics and outreach programs that teach responsible pet ownership.
However, knowing where to target limited resources can be difficult, according to Greg Miller, GIS analyst for the ASPCA. "The ASPCA used analytics to study the communities its shelters serve, but couldn't always uncover why, despite the availability of critical services, there hasn't been a reduction in incoming dogs and cats," Miller says.
The ASPCA is convinced the missing link has been location information, Miller says.
As part of a pilot program still in its early stages and funded by PetSmart Charities, the ASPCA identified three regions in the US - Cleveland, Portland and Southeast Florida -- to apply GIS-enhanced analytics using a combination of shelter management applications, the organization's own homegrown analytics software and Esri's ArcGIS.
"Data about species, age, whether the animal has been spayed or neutered and address aren't completely telling. Overlay that on a map and suddenly you can clearly visualize specific neighborhoods that are hot spots," Miller says.
Without GIS, ASPCA workers and shelters have to know the area well enough to see patterns among addresses in static reports. GIS helps illuminate trouble zones where disproportionate intakes of animals are occurring, rendering outreach far more effective. Instead of blindly passing out brochures or parking a spay/neuter mobile van in a random area, ASPCA and shelter representatives can blanket exact neighborhoods and place spay/neuter clinics in the most in-need areas.
The pilot project should save money, enabling funding to extend to services such as fences for dog owners, breed-specific training such as for pit bull owners and more in targeted locations, Miller says.
Like traditional analytics projects, the program's outcome is dependent on the quality of the data, which has led to new processes. Shelters now verify all intake addresses for the data to be considered clean. Also, they enter the location where the animal was picked up or seen vs. the address of the person who brought him in or called. Such accuracy, the ASPCA believes, will inevitably save animals' lives.
Making servers go 'tilt'
Adding GIS to business analytics is not without cautions, though, as it can overburden servers. For instance, EDENS plans to support mobile analysis of almost 200 layers of data from internal and external databases.
A data 'layer' consists of one theme, Beitz explains, such as a property layer for EDENS shopping centers and a property layer for centers that belong to other firms. Each of these layers contains thousands of data points.
While the ArcGIS SDE resides at headquarters, the mapping services that feed into it are in the cloud. Many of the mapping services are hosted in the cloud by Esri and are then fed into the application. EDENS' proprietary data is served as local map services, and they are hosted on the firm's on-premises ArcGIS Server.
Tapping into Esri's cloud servers was an application design choice driven mostly by cost. "At this point for us, it's cheaper to use the Esri Business Analyst Online API and make calls to Esri's cloud servers than it is to buy Business Analyst Server and host all the data ourselves," Beitz says. He pre-builds the most popular requests and models to speed response times.
Factors behind the GIS/BI marriage
EDENS is in good company in realizing GIS's ability to soup up business analytics. "GIS tools have existed in parallel with business analytics software until recently. Now we're seeing them merge," says Dan Vesset, program vice president for business analytics at IDC Research. IDC predicts a significant spike in the worldwide spatial information management market, which is how it categorizes GIS-enhanced analytics, from $3.1 billion in 2011 to a projected $4.1 billion in 2015.
Vesset attributes the interest in GIS-enabled analytics databases - including those from Oracle and IBM Cognos, and analytics-enabled GIS servers such as Esri's ArcGIS Server - to the proliferation and consumerization of GPS and GIS. For example, smartphones have geo-location capabilities that enable easy spatial data collection for back-end applications, and Google Maps and Microsoft Bing have reduced the high cost and complexity of GIS mapping software.
"Social media, vehicles, infrastructure - everything has sensors that gather geo-location data, and that has created a snowball effect," Vesset says.
The heightened availability of geo-location data has enterprises asking their analytics and GIS vendors to either build in blended support or offer it as an add-on, according to Vesset. And the GIS-enhanced attraction spans vertical industries.
For example, trucking companies use GIS-enhanced analytics to track and monitor waste such as fuel and idle time, local police departments use the technology to model optimum locations for crowd coverage for large events based on available personnel, and retailers use it to push targeted ads to shoppers' smartphones as they pass in front of their store inside a mall, Vesset says. All of these actions depend on spatial information paired with back-end analytics engines.
Improving decision-making at the EPA
The US Environmental Protection Agency (EPA) also has found GIS-enabled analytics to be a time-saver. The organization, which acts as an aggregator for government, academic, private industry and legacy databases, has started to count on the GIS/analytics combo to hasten and improve decision-making across all agency interests.
"We've united what have traditionally been two disparate technology stacks to carry out missions, evaluate policy and assess where we need to do additional work," says Jerry Johnston, EPA Geospatial Information Officer.
It has been more of an evolutionary process than an "event," Johnston explains, but there were a couple of drivers behind the marriage of the technologies at the agency. One was EPA's Environmental Information Exchange Network, an Internet-based system used by state, tribal and territorial partners to securely share environmental and health information with one another and EPA, according to Johnston.
The Environmental Protection Agency has "united what have traditionally been two disparate technology stacks to carry out missions, evaluate policy and assess where we need to do additional work," says Jerry Johnston, the EPA's geospatial information officer.
"Data from regulated industry has been coming in to EPA for years, and users naturally wanted to map much of this information," Johnston explains. "Uniting GIS with our Exchange Network has been a focus of the agency for some time."
More recently, to help manage environmental grants issued under the American Recovery and Reinvestment Act of 2009, the need to "bring together reporting information coming in from grant recipients with our geographic information systems was critical. "
These days, the EPA blends GIS and traditional data from myriad sources to monitor water quality across U.S. with its Hydrologic Benchmark Network. States, counties, municipalities and tribal nations continuously push geographically referenced and time-stamped stream water-quality updates onto the Network. If an anomaly is detected, the EPA can immediately tell from the integrated GIS data whether the station is upstream or downstream, which other areas could be affected and how soon. Because the data is displayed on a dynamic map, problems are immediately conveyed to anyone who needs to know, leading to faster response, Johnston says.
The EPA also uses GIS-enabled analytics for its ongoing air quality index. Pairing maps with current and historical air quality data as well as temperatures and wind readings guides operational responses. As one example, because asthmatic children have more symptoms when the air quality is poor, the EPA can instantly notify local officials so they can decide whether to close schools. Something as specific as knowing if the school serves a hilly or flat region, which GIS shows, would impact where the alert is issued, Johnston says.
The EPA combines its Oracle Database and Esri's ArcGIS Server with Google Maps and Microsoft Bing, simplifying otherwise complex tools for users. "Our users expect to see maps with their charts and models, so it's helpful that the analytics vendors are integrating GIS with their tools and the GIS vendors are integrating BI. That hasn't really happened before," Johnston says.
Though this consumerization is occurring at a rapid clip, some organizations, including the ASPCA, still rely on GIS experts to partner with data analysts for development, integration and analysis. In fact, the ASPCA's Miller, who has GIS Professional (GISP) and American Institute of Certified Planners (AICP) credentials, was brought into the organization specifically to help with the GIS pilot program.
IDC's Vesset agrees that adding GIS to the data gathering and input mix can complicate matters. "From a data management perspective, you need appropriate infrastructure and processes in place because you're dealing with large amounts of data," he says.
EDENS' Beitz has found the problem lies with data integration, not data entry. The cities and counties that his firm deals with have closed, proprietary databases, making it impossible to easily extract data.
"They don't have the capacity or desire to serve out their map data to be consumed as services in other applications," he says.
If the government data were to be widely available, companies would save a tremendous amount of time, according to Beitz. Today, to find information on an adjacent parcel of land where EDENS wants to expand, the company has to dig through the county's website.
"Imagine if we could open up our GIS application and display the counties' GIS data as an additional layer that can be queried," Beitz says. "We could set up alerts to tell us if the zoning has changed on any land within three miles of our property and track changes in property ownership near development sites."
VTN's Warren has experienced similar pains, causing the company's GIS database to be built bit by bit from data gathered during individual projects.
For EDENS, VTN, the EPA and other organizations, data openness is the next big frontier for fully integrating GIS with business analytics. "This would help us to more closely monitor our target markets and to quickly take advantage of emerging opportunities," Beitz says.