Search This Blog

Wednesday, June 27, 2012

Don't File It – Use That Data



Big Data is a means to extract useful information from great volumes of data collected from various inputs. If the data is available and accurate, the distilled information can be useful if the project is well conceived.

Last week we attended a Big Data session put on by NCA-GTUG to highlight some successful in-use projects with Splunk boiling down huge quantities of continuous real-time data into useful information about monitored activities – almost real-time. Fascinating.

Next day I participated in a GovLoop Webinar on Optimizing Service With Big Data, which outlined the expanding use of the Big Data approach in the Government sector – and an example of effective use of police data to reduce crime.

The city of Santa Cruz applied Big Data tools to 10 years of collected police crime data. They cut the data by type of crime; they sliced it by location; they diced it by time, day of week, day of month, and month of year, to get a detailed picture of the characteristics of crime occurring in Santa Cruz.

Then the process was reversed to create when, where, and what predictors of crime probabilities as a data based tool to apprehend or prevent area crime.

By manipulation of large volumes of existing data, the police department created a knowledgebase analogous to the instincts of long-service patrolmen with a 'feeling' about crime activity and targets. This predictive tool is intended to help the SCPD be more effective in the reduction of crime. Of course, personal experience is what sets the Doers apart from others – this data approach is a supplement, not a replacement for experience.

The results bear out this supposition: property-based crime is down by 12% and burglaries down by 25% since implementing the project. Good results, which supplement the other tools available to the police in doing their jobs – like when the two-way radios replaced the call boxes on the street.

This example demonstrates a valuable subsequent use of gathered data to glean information about patterns, trends, or other relationships among the measured items. Two important element to a successful project with massive volume of data is the relevancy of the project to get usable information and the completeness and accuracy of the input to get sound information.

What are some Big Data applications that could create valuable information for your operations or future planning?

1 comment:

Thoughthebrowser said...

I attended that Splunk demonstration where they were extracting value out of machine generated data. I've had to do that manually a number of times, and always found that I ended up doing it myself before I could coherently ask for the results required.
Once I had the process defined, it was still too labor intensive to attempt unless there was a big cookie coming at the end.
These tools allow us to extract more value from the data that is already gathered.