Business analytics is one of the top disruptive technologies today and it’s no wonder. Analytics platforms give organizations insights, so they can make better business predictions.

It’s also big business. Revenue generated from business analytics software is expected to reach $168.7 billion globally this year, up from $150.8 billion last year, according to Statista. And the International Institute for Analytics is predicting that analytics will be widely deployed by organizations this year to improve data.

Each year, Drexel University’s LeBow College of Business recognizes organizations that have achieved innovation in analytics. The Drexel LeBow Analytics 50 is a national recognition of industry analytics innovators where 50 companies are honored for their use of analytics to solve business challenges.

The Drexel LeBow Analytics 50 recognizes distinction in analytics across industries. From manufacturing and healthcare to government and retail, organizations are using analytics to improve decision making, combat crime and influence the way they operate and provide value.

The annual recognition reflects Drexel LeBow’s heightened focus at the intersection of academic and industry, creating a platform to share best practices in the field and to honor data-driven business impact. Organizations can nominate their analytics efforts until June 30 at

The judging panel is comprised of LeBow research faculty and industry officials. Nominations are judged by the complexity of the business challenge, the analytics solution implemented and the solution’s impact on the organization. This year’s honorees will be recognized at the third annual awards ceremony on November 1 at the university’s main campus in Philadelphia.

Here is a look at the analytics challenges and outcomes of five of the LeBow Analytics 50 winners for 2017.


One of the exciting promises of analytics is its use in identifying and quantifying health risks and making recommendations about interventions proactively. For years, the healthcare system has relied solely on limited physician experience, static patient models, conflicting therapeutic guidelines and often ineffective care coordination policies, maintains analytics platform firm IOMICS Corp.

In response, IOMICS has initiated a spinoff company called PrecisionCARE™, specifically for point-of-care analytics and personalized medicine.

Initially, We were really interested in pushing the boundaries on how predictive analytics could be integrated with the concepts of network medicine to accelerate cancer research,” says Joe Gormley, chief technology officer at IOMICS.

The idea was to develop a platform that would not only create a comprehensive view of this highly complex and multifaceted disease, based on the best available data, but to also analyze each patient and their condition and compare them to others across a spectrum of different discriminators and outcomes, he says. But as the tool was being built out, IOMICS officials realized the advanced modeling techniques they were using to integrate the broad range of patient data types with advanced machine learning could be applied to many diseases and potentially utilized beyond the laboratory.

IOMICS is currently running personalized medicine pilots. Officials are testing the power of the analytics software in areas including breast cancer risk and the consequences of non-adherence to care plans, advanced warning of stroke risk and even patient preferences to different therapies when all other factors are equal.

“There are a lot of variables that play into this type of analysis – from a patient’s current diagnosis to their co-morbidities, demographics and genomic profile,” explains Dan Corkill, chief science officer at IOMICS and past research scientist at University of Massachusetts.

The current healthcare paradigm may work well for 50 percent of patients, 50 percent of the time, “but the remaining patients are likely to be treated in a sub-optimal manner,’’ he says. “The goal is to provide a targeted and individualized decision about the recommended treatment for you.”

PrecisionCARE can automatically and continuously mine an institution's EHR system and supporting data to build a customized patient model on-demand, says Gormley. In traditional medicine, a physician may only see 20 people in his or her entire career whose disease profile looks like yours. Systems like PrecisionCare are meant to be an extension of a physician’s experience base and help them select an effective therapy.

The software, the patient and the healthcare provider should work synergistically, and in lock step, Gormley says.

“It’s not just a personalized model; it’s also how it is integrated into an EHR operational framework from which the healthcare worker can rapidly and confidently draft a care plan that can actually be implemented in a cost-effective manner," he says.

Corkill believes within five years, clinicians will be using healthcare analytics to help them make the best possible choice of existing and emerging treatments for a patient. “We’ve got to stop guessing about patient care and use a more rigorous scientific and data- driven approach.”

Seattle Police Department

In 2011, The U.S. Department of Justice investigated the Seattle Police Department over community concerns about the use of force and biased policing, which had culminated in the shooting death of a native American wood carver. A department officer resigned and was later terminated.

The SPD signed a consent decree in 2012, acknowledging it lacked the data and systems to support meaningful oversight by supervisors of their officers’ actions, and that front-line officers were not consistently supplied with the tools and knowledge to de-escalate instead of resorting to uses of force.

To aid its ability to demonstrate compliance with the provisions of the decree, the department rolled out a data analytics platform, “DAP,” in 2017, which provides new systems of oversight, risk management, and statistical reporting.

The DAP is focused on performance analytics: what officers are doing and critical areas that need to be analyzed, says Brian Maxey, chief operating officer at the SPD. Prior to the platform, “We didn’t have reliable data to demonstrate what we were doing.’’

The DAP draws from a variety of sources, including a computer-aided dispatch system and a records management system, he says. Officials also created a template for Terry stops, when a person is briefly stopped by police on suspicion of criminal activity. “We now record every single stop, and also every time someone is frisked,’’ Maxey says.

A crisis template was also developed, which records information every time an officer encounters someone they believe to be in some form of crisis that sometimes result in the use of force.

Since the DAP has been in place, “We found less than 2 percent of our encounters with people in crisis end up with any use of force,’’ he says. “This is the sort of granular data analysis the template allows us to do. We couldn’t have told you that four years ago.”

The DAP dashboards allow supervisors to see the data, which gets pushed out using visualization and analytics tools, he says.

Today, the SPD is making raw data available on its website for full transparency.

“The data is unprecedented in policing — all the information we gather about our Terry stops, frisks, crisis intervention and use of force, including officer-involved shootings—we put it out publicly,’’ Maxey says.

The department is about to release a supervisor dashboard to give sergeants metrics on every officer working for them. With ready access to how many calls officers are responding to, how many Terry stops, crisis calls and use of force incidents, “they’ll have a better situational view upfront,’’ he says.

Ideally, this will prompt sergeants to intervene or support their officers better because they’ll have more information, which should improve how officers are performing on the street, he says.

Police officials now also have visibility to detect trends and benchmark officers, bringing new understanding to how the department operates, Maxey says. “Because we have data around all these performance metrics, not only are we looking at adjusting policy, we’re looking at training and better ways to handle outcomes.”

Southwest Airlines

The cost of fuel is going up, and when you’re an airline with a fleet of over 700 planes serving 101 destinations, fuel consumption forecasting is critical. That’s why Southwest Airlines deployed an off-the-shelf system, Alteryx, to build an advanced analytics model to improve its ability to forecast and manage fuel inventory.

“Fuel is the number two cost from an operating expenditure perspective for an airline,” behind crew salaries, says Doug Gray, director of analytical data services, at Southwest.

When his team began looking at doing data analytics in 2015, fuel cost $50 per barrel of oil. Now it’s about $70 per barrel, he says. Southwest spends between $4 billion and $6 billion annually on fuel.

That prompted officials to start looking at how to optimize fuel purchases month by month, station by station, (which is airline parlance for airport), to see how much fuel was being purchased in Dallas versus Denver, for example.

When optimizing anything there are three things to look at: demand forecast; the purchase and holding cost; and a fuel shortage scenario, he says.

Previously, demand forecasting for each of Southwest’s over 100 airports was done on a rolling basis, 12 months a year, and all 1,200 forecasts were being developed manually in Excel.

“It took a person three days to do all 100 stations for 12 months, and the data wasn’t always accurate and wasn’t taking into account all the factors,’’ Gray recalls.

His team did a significant amount of custom development in Alteryx, which has a click-and-drag/drop user interface where a user can specify what data they are looking for within a database and ask the system to write an SQL query.

“Alteryx reduces the time to do what we call ‘data wrangling’ by about 60 percent,” he says.

Once the process was automated in the platform, the three-day, manual process to do fuel forecasting decreased to five minutes, Gray says. “So you can buy the right amount of fuel at the right station in the right month, every month. Now I know in every station, I’m forecasting correctly.”

This has resulted in “significant fuel demand forecast accuracy, which has led to significant tangible, economic and cost avoidance benefits,’’ he says. “We’re buying less fuel than we used to -- but not adversely affecting operations.”

Today, there are 600 Alteryx users at Southwest and analytics projects are underway in various business units across the company, he says. Currently, Gray’s team is using the same demand fuel forecasting inventory models and applying them to liquor purchases -- a big ticket item for airlines.

“It has a lot of the same characteristics as fuel,’’ Gray notes. There are 15 different types of liquor stocked on Southwest planes at 21 airports, and they are deploying Alteryx for more centralized liquor inventory management. “We want to get optimum purchasing power for liquor and free up cash flow and try to get a handle on how much liquor we need.”


A new analytics dashboard to better gauge how and where to grow soybeans began as the seed of an idea. That’s what Syngenta, a $12.65 billion global agrichemical company has built to help scientists select the best products and then help place the correct products on farmers’ fields.

Syngenta wanted to enhance the grower experience. For farmers, selecting which seed products to plant is an emotional decision, since that choice effects the success of their business — once a seed in planted, there’s no turning back.

“For decades, soybean plant breeding was considered an art,” explains Greg Doonan, head of novel algorithm advancement at Syngenta. “Plant breeders and farmers would rely on their intuition and what they observed to make decisions.”

Plant breeding is a complex and lengthy process. The development process starts with the selection of two soybean plants, called ‘parents,’ and cross-pollinating them to create offspring. Breeders then test the offspring for several years with the goal of selecting the best performing plants. The entire process can take up to eight years to generate a commercial product, called a variety.

“We wanted to give growers more confidence in their seed selection decision making,” says Doonan. “We first focused on a soybean plant’s development process.”

Rather than continue to rely on the old trial-and-error methods, where plant breeders simply looked at two parent plants to decide which two to breed, Syngenta has created tools to assist plant breeders during the development process.

“We wanted to make data driven decisions to improve the probability of successfully delivering a commercially-viable product,’’ Doonan says.

The process began in 2011. The analytics team developed mathematical models that move from what officials observed was happening with soybean breeding, to looking at different scenarios and predicting what could happen within different environmental conditions, he says.

The digital breeding platform the team developed contains genetic and geospatial data, such as what type of soil a farmer has, as well as historical weather data to gain a complete picture.

“There was a lot of hard work, especially on the data side, getting the data integrated, having data standards and understanding data was an asset we could make actionable decisions upon,’’ says Doonan.

The result of this work was Syngenta’s E-Luminate platform — a tool that Syngenta’s sales and agronomist representatives can use at the grower level. E-Luminate was implemented in 2017.

“We’ve improved our performance year after year, up to two to four bushels per acre, depending on the program,” says Doonan. “We’ve seen increased productivity in the breeding program, resulting in better performance in the field.”

The company has worked to “increase confidence of our growers’ decisions,’’ he adds, “and we’re known as a performance leader for soybeans with our growers and retailers that sell our product.”

The Dow Chemical Company

Improving how price, volume and cost are managed is a top business priority at The Dow Chemical Co., and executives were tasked with figuring out to use predictive analytics and optimize recommendations to enable decision making for its products.

Dow’s business and functional leadership used internal SAP data through custom reporting systems, dashboards and spreadsheets, which include current and historical data sets, to make decisions. While this met business needs, it did not allow for significant, breakthrough margin improvements.

A few years ago, the company migrated to SAP ECC – a project that took almost five years and was one of the largest SAP implementations in the world, says Puneet Sawhney, analytics value delivery leader at Dow.

“We have the advantage now of being on a single SAP instance globally, and this enabled our analytics team to draw data from a single source for all types of analytics from descriptive to predictive to prescriptive,” Sawhney says.

Part of the drive for predictive analytics was a mandate from Dow CEO Andrew Liveris, whose strategic vision was to transform Dow from “a commodity player to more of a specialty, value-added material science company,’’ says Sawhney. ‘The only way you can position yourself as value-added player is to truly understand what’s going on across your customer base, and that’s driven from insights data can provide.”

Because Dow had also acquired chemical manufacturer Rohm and Haas, there was a need for combined company descriptive analytics — reports, queries and visualizations -- so people in decision-making roles can look at data from a historical perspective, says Sawhney.

The analytics team created very robust reports that included key performance indicators (KPIs) and metrics, and a series of corporate dashboards, “so executives could get the pulse of what was happening in snapshots across the enterprise.”

The analytics team was also tasked with giving the business a forward view; meaning the ability to do diagnostic, predictive and prescriptive analytics.

For example, the team developed an order loading dashboard that shows a business user the pipeline of orders as they come in over the course of a month, so they can look at trends.

“Business leadership uses that to form a proxy to judge where the month’s sales results are going to end up being,’’ Sawhney explains. This applies to all materials sold across the globe, including polyurethanes, plastics and construction chemicals.

Today, there are close to 100 dashboards in use at Dow in various aspects of the business. Sawhney says the analytics team has been delivering “hundreds of millions of dollars in value in decision-making in top line revenue growth and expanded margins, and where to reduce costs and gain productivity.”

Other projects include forecasting work for management of demand plans and forecasting for price volume changes, and techniques to optimize Dow’s assets and determine what to produce, where to produce it and how much to produce globally.

“We use that data to maximize our profitability,’’ he says.

More recently, the analytics team has been working to develop a data as a service for business stakeholders. Using power tools, business stakeholders can mine data sets and make interpretations to gain insights.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access