How to operationalize advance data science at speed and scale

As agencies become more digital and connected, data boundaries blur providing agencies access to a variety of data sets in multiple formats and from multiple sources. This drives agencies to solve evolving social and population health issues, including the COVID-19 crisis, by adopting advance data-science based approaches that leverage AI/machine learning (ML) tools.

Innovation in cloud technologies, easy access to increasing computational power (CPU, GPU, Hadoop), intelligent statistical algorithms and powerful software are making it easier for agencies to adopt these advance data-science based approaches. In their survey, NASCIO found that over 55% of state IT organizations are pursuing AI initiatives and another 32% are running AI in some production operations or staging pilot projects. However, despite investing significant time, effort and resources, 87% of these AI/ML initiatives fail.

Complex analytics projects must be simplified by using AI to accelerate routine tasks and to democratize data-science skills enabling organizations to decentralize analytics, making it more accessible and self-service based.

An automated data-science approach can help agencies automate the foundational data management activities and the AI/ML journey, enabling citizen data scientists to build, train and deploy effective machine learning models across various program areas quickly -- arming policymakers with timely insights in an environment where real-world data is changing rapidly.

From democratizing data science skillsets (i.e. making it easier for anyone to build ML models) to accelerating the model deployment lifecycle, this approach can help public sector organizations effectively use AI to improve program outcomes and citizen experience.

The point of view discusses the approach and explains how agencies can adopt the same.