For successful data science projects, talent is not enough

By Ijeoma Pelecanos, Ph.D.,

Subscribe to our newsletter

Data science projects have rapidly grown in recent years as companies focus on uncovering new insights. If your company's data science projects are successful, your organization can be one of the insight-driven businesses projected to take $1.2 trillion each year by 2020.

Core to an insight-driven business is the data scientist — the creator of the analytical models used to derive critical insights from data, driving new and differentiated applications, services and business models. These skilled specialists are in high demand, and that demand far outpaces the supply.

Faced with this reality, CIOs have a critical role to play to ensure their businesses are positioned to succeed in their data science projects. Though talent acquisition is necessary, it's not enough. CIOs need to ensure their data scientists are highly effective in their core mission.

The first step is to understand the top challenges of data scientists. What are their concerns? What slows them down? Here are some of the hurdles they face:

Data accessibility

According to Michael Schmidt, data scientist and founder of Nutonian, "Gaining access to data and getting it into the proper format can be extremely time-consuming. ... And it's an area in which I'd love to see more tools to make [the process] easier and faster."

CIOs need to eliminate data silos to allow easy access to data across their businesses while implementing the right security and governance to ensure teams aren't misusing the data. The user-friendly tools they implement to aid and shorten the data preparation process, a time-consuming step, is also key.

Collaboration

"Changes are happening so fast that we have to rely on diverse input for us to work together and drive better business outcomes. ... It really helps to have that collaboration across domains, across subjects, across techniques and with other people," says Andrew Huynh, who was at the time a data engineer at Funding Circle. Fostering collaboration between data science teams and other domains to speed the use of insight to create new applications, services and customer experiences is also vital to success in data science projects. This collaboration requires a feedback process to continuously refine analytics models and algorithms based on domain expertise to ensure insights are always relevant in the rapidly changing market.

Cutting-edge technology

"Data scientists are inquisitive. ... They will always be learning and thinking about what new technologies are out there that will help them be efficient and help the business succeed," says Benjamin Skrainka, formerly the principal data scientist at Galvanize.

There are several valuable open-source tools from the data science ecosystem, with new innovations continuously rolling out. Integrating these cutting-edge technologies to ensure data scientists can use the best available technologies while ensuring security and deployment stability will have a significant impact on their success.

If CIOs take the initiative to address these challenges, they will enable their data scientists to rapidly uncover new insights and collaborate across domains to transform that insight into action in the pursuit of disruptive new business opportunities. In doing so, their data science projects will succeed, their businesses will become insight-driven and the strong business results that companies such as RSG Media achieve can become a reality.

How RSG Media empowers its data scientists

RSG Media, a global software company, designs and develops media management analytics technologies to help its customers optimize programming schedules and deliver hyper-targeted audiences for their advertising and marketing campaigns.

To address the core challenges of its data scientists, RSG Media invested in the following:

  • Improved technology to manage and prepare data, reducing the data preparation labor cost from eight people to one person
  • An enterprise-grade platform with the critical security, governance and scalability delivered as a managed service with integrated, state-of-the-art and open-source capabilities, such as Apache Spark for rapid data ingestion and algorithmic processing
  • A workbench that provides data scientists with a learning and experimentation environment to support a hackathon-style culture of rapid, collaborative development to speed time to market

By empowering its data scientists, RSG has achieved significant results: One client saw a $50 million lift in bottom line through advertising yield operation. In addition, RSG saw an average of 40 percent lift in reach, for a 24 percent boost in new viewers. This is the kind of transformation that's possible when data scientists have the right tools, teams and access to data.

Recommended articles

Watson
I'm still learning and don't have a suggestion yet. Stay tuned.