Page 29 - FCW, October 2017
P. 29

the things we need to do. This is driving a liberation
of analysts who have been stuck with subpar tools. We now have an opportunity to reinvigorate the analytical workforce within the agency by equipping them with tools that work faster.
“We now have an opportunity to reinvigorate the analytical workforce within the agency by equipping them with tools that work faster.”
What does that mean for the kind of resources in which you need to invest?
It’s about the capacity of our people to do the work, and ensuring we equip them with the right tools to do this kind work safely, securely, and ethically. There needs to be a data management and governance process underlying all of that
with the aim of providing those analysts with the best possible data. That means having programs and practices in place not only to improve the quality of the data where it lives, but also to democratize the availability of that data.
A big concern is having access to data that is both clean and reliable. What improvements can and need to be made?
This is indeed a challenge, and it breaks down into two kinds of dynamics. Take the census of fatal car crashes. That begins with a police crash report, and that’s a very human endeavor. It’s open to a great many challenges as it relates to accurately recording data and ensuring the data fields are complete and consistently filled out. Still, when there are inconsistencies
and things missing in the data, we have to deal with that. Our systems and our analytics must be resilient enough to deal with human frailties and the needs of the real world. That’s never going to go away.
How do you factor this less-than-perfect data reliability into the analyses you have to do?
There are 60,000 people in this department. Our analysis and analytical expertise are distributed throughout 10 operating administrations and across the country. My job is to give them all the tools they need to do their analyses. I don’t decide whether any given data is fit-for-purpose.
And that’s OK. What I want to do is build community, to share code, and make it easier for people to find folks working on similar problems so they can share their practices and technology. We need to continue to build that culture into the agency so we can overcome the challenges. To that extent, I’m more of a trust broker than anything else.
What does all of this mean for system performance and security?
There’s a bunch of data coming off those systems that we could use to tune their performance and improve security. Then there’s the way we can architect our systems to better feed the analytical needs of the agency. That starts to point to ways we need to work through our system architectures
to distribute the processing workload. That means thinking about cloud architectures that we can use to help us offload some of the processing from our network and servers.
What are some emerging technologies that could make a difference to how you manage and deploy data?
There’s not much that is truly brand new. And it’s not just about the technologies themselves, but the user experience around them that is important. For example, I’m looking for ways to make it easier and less expensive to build an analytics platform using a server-less architecture, and I’m starting to see cloud providers understand some of that.
For the most part, I believe it’s about the evolution of things already in place. Technologies that make it simpler, easier, safer, and more secure to do things with data are definitely where I’m looking for the most advancement.
SPONSORED CONTENT
















































































   27   28   29   30   31