DataOps Can Move Analytics from a Discrete Activity to a Winning Business Discipline
The potential of artificial intelligence (AI) technologies and the growing shortage of data scientists are pushing financial institutions to consider alternative approaches to their data science teams and data analysis practice. A winning strategy must bring big data analysis together and make data science work visible, shareable, reproducible, and standardized. A key enabler of this is the introduction of DataOps and teams that support a variety of roles across the data processing pipeline.
My recent report examines the use of DataOps and supporting tools across the data science workflow to help FIs automate and expedite the tasks of developing and running analytic models. Celent believes DataOps has the potential to industrialize data science, through improved repeatability of findings and reduced time to identifying actionable insights.
It describes the context of data science and AI, provides insight into the process for creating and deploying an AI model, and provides details about DataOps and its impact for FIs desiring to make their data initiatives more efficient. It also provides a helpful categorization of tools FIs can use when launching AI-based initiatives. It answers the following key research questions:
- How can DataOps action data science?
- What benefits can DataOps bring?
- How do FIs of all sizes reap the full benefits of data science?
Celent defines DataOps and Data science as following:
DataOps: A set of processes and tools aimed at expediting the creation of data and AI products, from development to putting the models into a live environment.
Data Science: An interdisciplinary field employing mathematics, statistics, machine learning, and computing to extract knowledge and insights from data.