Data Science for the Masses

Once exclusive to large financial and technology companies, data scientists are now finding homes in organizations of all types and sizes, helping to extract invaluable insights from the ever-increasing data deluge that every business is facing. This article looks at the expanding roles of data scientists and offers practical advice on performance-enhancing technology choices.

Thanks to the growing information deluge, organizations of all types are hiring data scientists for a range of roles, from predicting future revenue to improving the quality of a customer’s experience while visiting a website or visiting a retail location. The volume of data from IoT sensors alone brings new ways of looking at industrial, transportation and logistics, helping companies find and fix bottlenecks, increase efficiency, and track and reduce defects.

For example, retailers can personalize the shopping experience and tailor product offerings to meet anticipated sales. Uber uses data science to maximize revenue by adjusting pricing to demand. And content organizations such as Buzzfeed use data science to track and adjust articles based on engagement, to help maximize views and clicks for their advertisers.

Turning data into action requires a number of tools—machine learning, modeling and artificial intelligence, to name a few—and data scientists often use compute-hungry software such as TensorFlow, PyTorch and MATLAB to get the results their organizations demand. However, it takes more than software to improve the performance of data scientists and their applications.

First, understanding what the data means is critical to a project’s success. Surprisingly, in a November 2021 study by HP, 40% of data scientists said they often start manipulating data sets before they have a good understanding of the business objectives for their project.¹ Solid communication between data science teams and the ultimate users of the data will help ensure projects start and stay on the right track.

Since many data scientists use both Linux- and Windows-based tools, a real productivity boost can be achieved by using a workstation that supports Windows Subsystem for Linux, which lets users run Linux tools and applications directly from Windows without requiring dual boot or virtual machines. Productivity can be further enhanced by using systems with preconfigured software stacks designed for data science, such as Z by HP workstations, which include comprehensive suites of application and environments that all but eliminate software incompatibilities. In the HP study, 42% of data scientists said they lost an average of five hours a week just tinkering with their data environments—time that could be better spent creating and training models.

Workflow optimization is critical to any task, and data science is no exception. Although many users have developed their own optimization strategies, taking advantage of new optimization tools—especially workflow automations—can make a big difference in productivity, especially for those tasks that require a lot of repetitious, manual processing. Automating those tasks will save time and energy, enabling entire processes to run autonomously with just a few mouse clicks.

There are many ways to improve the performance of data scientists and their applications, although the power of the user workstation may be the biggest factor in delivering top performance and productivity. Z by HP is a comprehensive line of workstations designed and preconfigured for data science applications, with tools and libraries preinstalled to ease day-to-day activities, save time and keep users productive.

Click here to learn more about how Z by HP is constantly evolving to meet the ever-changing demands of data science.

¹ “Understanding Data Scientists,” HP proprietary research, November 2021