Blog
Democratize and Scale Data Science with Citizen Data Scientists
Thomas Haslam, Director, Product Management, Insights Generation
Jul 19, 2021

As the talent shortage of formally trained data scientists continues, it is more critical than ever to democratize data science by empowering business users to refine and optimize algorithms without delving into code.

Non-technical individuals who use data science tools to solve business problems are commonly referred to as ‘Citizen Data Scientists’. They provide the industry experience and local business knowledge that a data scientist may lack but need support with more technical components. One option is to provide algorithm templates that can be reused and tweaked for different use cases.

Organizations realize that the key to scaling data science is to increase collaboration between data scientists and citizen data scientists. When the technical skills of data scientists are combined with the domain knowledge of business users, the result is reduced effort and increased efficiency. Historically, the common approach was to upskill data scientists to understand business problems. With the rise of citizen data scientists, the opposite is true where business users are equipped with extendable machine learning and the ability to customize parameters in algorithms for various needs.

IQVIA now offers new insight-driven omnichannel accelerators, including our Customer 360 Feature Designer and Algorithm Studio, that help productize machine learning algorithms for faster scaling to multiple therapy areas, brands, and markets.

Customer 360 Feature Designer acts as the interface between individual algorithms and the data they run on: collating all information from the data warehouse or real-time pipelines and deploying to each model as required. It removes the need for individual model pipelines per algorithm removing duplicated effort and provides visibility into data lineage and governance while also making algorithms easier to share and reuse across teams.

Algorithm Studio provides templated pharma-specific algorithms that enable greater flexibility, for example allowing business users to easily change the number of channels included in a Next Best Sequence algorithm. It also allows configurability for specific parameters within each algorithm to fine-tune deployment.

These components help overcome common challenges of achieving algorithm extendibility, such as data variation by market and localized business rule requirements. Our vision is to close the gap between business users and ML engineers by encouraging the co-creation of initial algorithm shells, which can then be deployed entirely by the business user within Algorithm Studio. Closing this gap allows faster scaling, but also ensures the expertise of both data scientists and business users is leveraged to maximize success.

Productized machine learning is a giant leap towards empowering business users and managing the workload of data scientists with the end goal of scaling and democratizing data science across the organization. Explore the Orchestrated Analytics webpage to learn more.

Related solutions

Contact Us