For the Bureau of Labor Statistics, the push to modernize and consolidate its technology infrastructure is kicking into high gear.

BLS has been trying to reduce its servers, applications and other IT components for several years, but for the rest of 2016 and into 2017 that effort will come to fruition.

Carol Mullins, the associate commissioner for the Office of Technology and Survey Processing for the Labor Department’s Bureau of Labor Statistics, said the end goal is more centralization of back office servers and functions.

She said, for example, BLS consolidated all of its Oracle databases in one location, created a central storage facility, and brought several common tools together into one data center.

“That’s allowing us to reduce cost, gain some efficiencies, allows us to leverage the skills of our staff much better and maximize those skills,” Mullins said during her interview on Ask the CIO. “We, of course, also are focused on data. BLS is a data organization, and so we are continually looking at our infrastructure to ensure we have sufficient capacity to store the data that the BLS programs are collecting. There’s a lot of work in the bureau to look at alternative sources of data, big data and data sets available in the public domain so making sure that the BLS infrastructure is sized sufficiently to support that big data coming into the bureau is another big focus.”

The big data challenge is one that is at the center of Mullins priorities. She said annually BLS reviews its central storage facility to understand usage and potential capacity for the next year.

“In order to get that kind of information, we have to work closely with the business side of the bureau to understand what they are thinking about and working on in the up-coming fiscal year to make sure we are sized correctly,” she said.

The communication with the business side goes even further when BLS is trying to ease the burden of collecting and analyzing data.

Mullins said the bureau is testing out text readers to improve the speed at which data is coded as well as to reduce human error.

“Text reading requires tools and approaches that haven’t been used before in the bureau so that requires IT support both from an infrastructure perspective and from an application development side building systems that can take that data and use it for purposes throughout the bureau,” she said. “Some of the data the bureau gets is a five or six word description of an accident or injury that has occurred to an individual. Rather than have somebody read that and go through a big long list of codes and try to figure out what code that five word description matches to, we are trying to use the computer to read those words. And by process of learning, code that description with the correct code and, of course, it takes time. People sometimes refer to this as machine learning.”

Mullins said BLS is applying software to its applications and building models to test out the machine coding concept. Then a BLS employee compares the computer’s coding to a human-coded effort to gauge the accuracy of the machine learning effort.

“I don’t know if we will ever get to 100 percent for some of the areas where this research has really been focused. I believe the machine accuracy is up around 95 percent, so it’s quite high. We are still in the early stages of it and there is quite a bit more work to be done. We are looking at other parts of the bureau were we are not currently using this approach to see if it’s applicable too.”

Mullins’ office also is working with the business side to better understand their needs about bringing in new sources of data, which may require special handling or special security.

The bureau’s focus on big data isn’t all internal. A large portion of the information the BLS collects is out on its website from the job rate, to the consumer price index to a host of other economic indicators.

Mullins said BLS developed an Application Programming Interface (API) and a data finder tool to help users discover the information they are looking for.

“We are putting more of our data out there on the BLS public website for whoever wants to use it. The initiatives that have come out recently really haven’t changed our approach. The commissioner is very keen on maximizing the data that is out there and making it easy and available for people to use. We offer it in multiple different formats,” she said. “We’ve always been an organization full of data geeks and data nerds so the recent push to put more data out there and make it accessible is really just welcoming more organizations and agencies into the world that we’ve been in for some time.”

BLS’s consolidation and centralization push also is supported by the implementation of the Federal IT Acquisition Reform Act (FITARA) law.

While Labor struggled with how best to meet the requirements under FITARA, Mullins said her office is engaged with the department’s chief information officer around IT budget and project planning.

“Some of the things that were embedded in FITARA were requirements or activities that were already in place coming from the department. For example, every year I send over to Dawn [Leaf], the CIO, a list of all of the things we intend to spend money on at the bureau that is tied to IT and Dawn approves that big, long list. That is a critical piece of FITARA that was already well established prior to FITARA,” she said. “Some of the other pieces of FITARA BLS has had some concerns with, the federal CIO [Tony Scott] did send out a memo providing additional guidance to agencies that have a statistical agency embedded in them. So, we are hopeful that guidance will be a way for us to work closely with Dawn and her team as we move down the FITARA path.”

Mullins said the monthly meeting of the enterprise implementation committee (EIC) is one good place where the collaboration and communication around technology occurs. She said Labor is leading an effort for unified communications initiative as well as software licensing.

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.