Search: Specialized strategies yield results

Search applications with complex data models pose special challenges. One such example was a search application Contegra built for Carrier Corporation. The number of parts is large, and many parts have revisions to them that have been built over the years. Serial numbers, model numbers and other identifiers must be incorporated. Also, the system had to be able to identify what product the customer originally received because that would affect the replacement part needed. In addition to searching the database, the application must be able to locate technical documentation, brochures and other files in a variety of formats.

Perspective is key

“This project required a lot of complex design,” says Wiesenberg. “DtSearch is able to search across SQL databases and PDF documents, and then search parts referenced within documents. It sounds like a simple task to recognize parts numbers, but when you have 50,000 of them with complex relationships and want to do the search quickly, using the right search tool is very important.”

Understanding the search needs of different users was also important. “Engineers see the world as a set of different systems, such as engines and drives systems, which have specific code numbers,” Wiesenberg says. “There are hierarchical taxonomies, model numbers and serial numbers. Retrieving the exact information for each installation is critical when considering the action being taken is a certain type of repair or service.” Each user group could have a different perspective on the data that requires it to be searched and viewed differently.

Particularly in specialized applications, the upfront work is critical. “We want to be sure that when the users do a keyword search, the information is presented in a meaningful context, so they can narrow it down to the exact part and the nature of the document, whether it is a technical publication, leaflet or catalog,” Wiesenberg explains. “A good content model enables both the field service staff and external customers to access information in a self-service mode, which cuts costs for the company.

Geospatial search

Geospatial information has become increasingly important for many different applications and analyses ranging from marketing to agriculture. Yet the management of geospatial information has lagged that of any other kind of information. Although maps and imagery can be stored in a repository like any other digital files and searched according to indexed metadata, the ability to perform more complex searches on the data and process it once retrieved has been limited.

Voyager Search is a search solution designed specifically to manage spatial information. “We took modern search technologies and combined them with a unique understanding for geospatial data,” says Brian Goldin, co-founder and CEO of Voyager Search.

First, the company developed an indexing solution that could extract information from nearly any type of geospatial data regardless of format. It later began supporting other non-spatial documents such as PDF and various Microsoft Office formats. “Now we are combining the two to offer a solution that enriches data through linking documents to a map,” adds Goldin.

Users are able to define a geographical area on a map and then search for relevant information about it. “The user can find all the river data or stream flow data in that area, for example, or reports by dragging a box on a map,” Goldin continues. “This ability is not available in other search engines.”

Voyager Search can manage very large quantities of spatial data. An agency in the intelligence community chose Voyager Search to index several billion documents. Making data available and accessible while keeping it secure is a big ?challenge. A geospatially enabled search solution is critical not only to providing online access to geospatial intelligence, but also to broadening the analytical expertise of the ?organization. Voyager accomplishes that by providing the tools to index a wide variety of content and to add a layer of geospatial intelligence to the index.

The process for narrowing results from a search of geospatial data differs from that of other types of data. Search results from a demographics database, for example, might be narrowed by selecting only one income category; in a spatial search, “refine your search” might mean “view only content in a specific geography” in order to look at one aspect of the results.

Voyager Search can also create new files derived from geospatial and image data. For example, a petroleum engineer could look for seismic shockwave data from a certain time interval for use in his or her analysis. Emergency responders might look for recent areal photographs taken in their area and run them through an imagery analysis process to look for hot spots that would indicate the most recent fire perimeter. Voyager Search offers that capability, known as “geoprocessing,” which includes a variety of ways to manipulate geospatial data at the desktop.