CUTTING BIG DATA DOWN TO SIZE

IxReveal’s uReveal analytic software employs four algorithmic methods to derive more meaningful results from structured and unstructured data.

UNBIASED LEARNING - Without any query, analysts can discover meaningful concepts. Addresses the concept that “you don’t know what you don’t know.”

OLAP FOR TEXT/ COLLABORATION - Analyzes text like spreadsheets on numbers. Regardless of numbers or codes, analyst can mix, match and build concepts like LEGO blocks and share expertise.

AUTOMATED CLASSIFICATION (context disambiguation) - Automatically classifies flu bug, software bug, insect bug, spy bug and Volkswagen bug.

RELATIONSHIP DISCOVERY - Relates disparate mentions of concepts such as “automobile was going fast” with “car was speeding.”

Source: IxReveal

Three years ago, the crime analysis unit of a mid-Atlantic city police department wanted a way to track drug and gang activity by pulling information from police case reports and other sources found in disparate files and data sets. The unit, which asked not to be identified, has an in-house database linked to the department’s records management system, but the database has search limitations and is not very intuitive.

The unit turned to IxReveal, an analytics company whose software lets users fuse and extract trends from any data, structured or unstructured. The software — uReveal — “harmonizes” data by fusing information without the need for extensive data-mapping, according to Ren Mohan, IxReveal’s co-chairman and CTO.

“Our platform bypasses extraction by empowering any user to identify entities or concepts of interest in data in its natural state,” Mohan said. This is a significant shift, he said, providing a layman-friendly approach that increases speed of processing and cuts IT costs.

For instance, a police analyst might be looking for information about a potential suspect who is a gang member, carries a silver gun and has fought with the police. “Silver hand gun,” “gang member,” and “hostility towards police” would all be concepts of interest, according to an example presented by IxReveal. The analyst would then select the data sources where this information could be found — databases of police or arrest records, crime incident databases, tips from the public in PDF format, RSS news feeds or Web pages.

The software then allows the analyst to use link discovery analysis to discover contextual relationships in thousands of documents. In the example given, the suspect, who carries a Desert Eagle chrome-plated silver hand gun, was discovered not in the weapons or property fields of the document, but in the police officer’s written narrative.

The city's crime analysis unit is using the software to create concepts, which can contain different search keywords, phrases and exclusion words to determine if any fall within their data sets. “You can take structured reports and fields and turn them into unstructured [data], so you can search phrases against narratives or other fields in that report,” an officer from the unit explained. “We use it a lot for tracking gang activities, drugs cases and drug trends.”

For example, a specific gang such as the transnational Mara Salvatrucha gang might be mentioned in various reports as MS-13, MS13 or Mara Salvatrucha. Police analysts can lump all the information together, extract pertinent case information and use other software programs to manipulate data, an officer explained.

A list of cases can also be mapped into ArchMap, a component of ESRI's ArcGIS suite of geospatial processing programs used to view, edit, create and analyze geospatial data. Or the information can be imported into a visual intelligence tool such as IBM’s i2 Analyst’s Notebook.

Using all of these tools, police can create link charts, point maps and hotspot maps to show historical areas that have gang activity or specific types of drug activity, which is helpful to patrol officers or gang and narcotics detectives, according to the officer.