Most agencies were built as extremely centralized organizations, where policy and direction trickle down from headquarters and data and results circle back around.
COVID-19 brought a shock to that system. Agencies converted to telework, and newly distributed structures saw data exclusively generated, processed and analyzed remotely.
“The decentralization of agencies exacerbates the drawback inherent to the centralized system,” said Harrison Murphy, Director of Data Analytics Solutions for Altair, a global data analytics software provider for government.
Agencies have no choice. They must adapt to an environment where employees can work with data from anywhere. When agencies fail to adapt, they risk bottlenecks in their processes.
In an interview with GovLoop, Murphy laid out three areas agencies should focus on.
Data used to be generated through highly controlled organizational processes, such as logging or filing constituent information. Now, data creation is increasingly driven by individuals.
The keys to success in remote environments are flexibility and agility. Agencies need clear strategies and governance policies that dictate how data should be acquired, prepared and structured, so employees aren’t left stranded without guidance. These groups also need systems to place controls on data while still giving employees the freedom to innovate.
Remote work realities mean employees can’t peek over a cubicle and ask for help. Despite this barrier, collaboration and security are increasingly important as insights are locked away on individuals’ computers. Agencies need a platform that allows users to work on the same datasets at the same time, while also eliminating the security risks associated with housing data on employees’ laptops.
“We see collaboration going beyond the simple passing of the data baton from one user to another,” Murphy said.
In the office, a data analyst could share a multifunctional model on a hard drive. Those same capabilities are crucial remotely too, but they must be available in the cloud.
Data acquisition and preparation are critical components of self-service analysis, which unlocks invaluable information, gathers insights and identifies trends. In today’s era of remote work, without the right tools, agencies’ scales can tip even more in favor of preparation instead of analysis.
Non-tabular sources, or language-based data types such as online forms, often need to be coded by data scientists – unless agencies have a solution that removes the heavy lifting.
“Agencies need to find the balance between allowing end users agility and flexibility but still maintaining a governed, curated infrastructure,” Murphy said.
With Altair, agencies’ structured and multi-structured non-tabular data can be standardized and prepared for analysis in a user-friendly platform online. Then, data analysts can generate supervised and unsupervised models in the cloud – in a central location even though they’re not in the office.
The final step unlocks visualizations and advanced analytics. But none of that’s possible without the hard work of adapting processes to a distributed environment.
“Rather than trying to reestablish ‘the way things used to be,’ agencies may gain efficiencies by addressing the pain points of current data processes, allowing them to flex with future interruptions,” Murphy said.