Data Science: Future or Fiction?
Data science is hailed by some as the key to business growth and success. But is its role overstated?
Smart companies collect and analyze a wide array of information related to their products, services, customers, supply chains, production lines and more to obtain deep insights into sales, marketing and operations. Other companies, however, waste much of their data, despite amassing it from a growing number of sources ranging from online sales to shipment tracking systems to social networks. This potentially insightful information is left unused simply because its value was never recognized or the company didn’t understand how to approach it.
Yet while many organizations are wasting data, pioneering businesses are determined to squeeze every last drop of value from each bit of information they collect. These companies use a science—data science—to leverage their data into actionable intelligence.
Data science delivers unique business benefits that can’t be found elsewhere. It has the power to unlock information that leads to sizeable revenue gains and long-term business success. Most organizations have come to the conclusion that the future belongs to those that know how to collect, manage and extract value from their data.
William Rand, assistant professor of marketing, decision, operations and information technology at the University of Maryland and director of the school’s Center for Complexity in Business, says that leveraging data is inevitable in today’s competitive environment. Organizations must embrace their data or face serious consequences.
“Data needs to be organized into information and then transformed into knowledge to become useful for managerial application.”
—William Rand, assistant professor of marketing, decision, operations and information technology at the University of Maryland and director of the school’s Center for Complexity in Business
Rand notes that many businesses are throwing away potentially useful data because they don’t see its value or they lack the capability to capture and analyze it. “In a recent survey, we found that although there were a lot of firms that were collecting large-scale data sets, [many] did not have the skill set to transform that data into something useful, Rand explains. Data science, and the experts who understand how to apply its principles, are now helping businesses fill that knowledge gap.
Data science differs from related disciplines such as statistics, data engineering, pattern recognition/learning, visualization and uncertainty modeling, because it uses all available and relevant data to present insights and forecasts that can be readily understood by business executives and managers. Data science’s popularity has soared in lockstep with the growing awareness of the potential knowledge tucked away inside diverse data types.
Given the ability of data science to put all types of data, including multi-structured data, to work for the business, many experts believe that data science isn’t a passing trend and will remain relevant into the foreseeable future. “It’s here for quite a while and it’s here to stay,” says Vincent Dell’Anno, the Denver-based global big data practice lead for Accenture.
“New tools, new technologies allow me to do things more quickly, but the overall goal and the things I'm trying to solve with data are essentially the same.”
—Vincent Dell’Anno, the Denver-based global big data practice lead for Accenture
Unique Business Value
Data science helps businesses tap into insights collected from all data, including social media, sensors and other emerging data sources. It spots trends, shifts and patterns that would otherwise go unrecognized—and unused. The technology builds upon traditional data mining by adding social media and additional types of unstructured data into the analysis pool. Data science then provides deeper insights into markets, supply chains and other key business areas.
Diego Klabjan, associate professor of industrial engineering and management sciences at Northwestern University, points out that data science can add business value by tracking the opinions of social media users in the immediate aftermath of a new product launch. “What used to be a lengthy process to gauge sales data can now be performed in real time by establishing the sentiment of customers on forums, Facebook and other social outlets,” he says. If a preponderance of negative comments is detected, the business can take quick remedial action. “The company can, in a short period of time, pull the product from the market or redesign it in case of prevailing negative comments,” Klabjan explains. “On the other hand, a positive tone gives the signal to ramp up the production and distribution.”
“Mashups combining various data sources will become a ubiquitous part of many data-driven solutions.”
—Diego Klabjan, associate professor of industrial engineering and management sciences at Northwestern University
At the same time, innovative data mashups can help executives gain deeper insights into various business challenges. “Data from online media and call center conversations is hard to store and analyze in a centralized database,” Klabjan says. However, substantial value can be obtained by using data science to merge data from traditional sources with new streams, such as online media and call centers. “Mashups combining various data sources will become a ubiquitous part of many data-driven solutions,” Klabjan predicts.
Another way data science can help businesses is by creating models that executives can easily modify to predict outcomes. “At the Center for Complexity in Business, we often employ the techniques of agent-based modeling (ABM) and social network analysis (SNA),” Rand says.
ABM provides the ability to model consumers, competitors and others in a particular market, while SNA enables the modeling of social relations and links of influence between individuals in a realistic way. ”Because such tools represent concepts that management is already familiar with, it makes it much easier to interpret the results,“ Rand points out.
Actualize Data Science
Most companies getting started with data science hire a data scientist to lead and coordinate the research, and relay key insights to executives. “A data scientist should be able to identify the right data sources to meet the objective and the entire back-end process should be transparent to the management,” Klabjan says. Storage, analysis and visualization tools, meanwhile, provide the foundation for a data scientist’s toolkit.
Dell’Anno notes that software is playing an increasingly important role in helping businesses manage and manipulate data that is now pouring in from multiple sources. “As scale and complexity of the data start increasing, you can’t do it all manually,” he says. “You have to take different approaches.”
“A good data scientist should be able to give you a probability—’we think this is likely to be correct in 95 percent of the cases, 92 percent of the cases’ and so on.”
—Mark Whitehorn, analytics chair at the University of Dundee
He also observes that powerful and sophisticated data science tools allow age-old questions such as, “How can I retain more of my customers?” to be answered faster and with better accuracy. “New tools, new technologies allow me to do things more quickly, but the overall goal and the things I’m trying to solve with data science are essentially the same,” Dell’Anno says.
Still, while data science can benefit businesses of all types and sizes, executives must be careful to not let high expectations outstrip reality. “Can data scientists guarantee that the answers you get are accurate? No, we’re not in that territory,” says Mark Whitehorn, who holds the analytics chair at the University of Dundee. “Of course, a good data scientist should be able to give you a probability—’we think this is likely to be correct in 95 percent of the cases, 92 percent of the cases’ and so on,” he says.
Make Something Out of Nothing
Data is, by itself, useless, according to Rand. “Data needs to be organized into information and then transformed into knowledge to become useful for managerial application.” That’s the role data science is now playing at a growing number of businesses.
While collecting and storing data is no longer a pressing business problem thanks to data warehousing, squeezing insight from an expanding torrent of information is now a major challenge for any organization. By tapping into customer views, analyzing data from multiple sources and enabling executives to explore data in imaginative ways, data science turns information into a powerful business tool.