Predictive analytics is a statistical technique used to analyze current and historical data in order to make a reasonable prediction about future. In a business environment, organizations employ predictive analytics model to identify market trends, opportunities and risks. Using the predictive analytics, organizations are able to assess potential risks and opportunities to achieve competitive market advantages. In other word, predictive analytics is part of data mining focusing on extracting information from historical data and used the data to predict behavioral patterns and trends. Typically, predictive analytics can be applied to any type of unknown events in order to predict the presents and future events. Banks are the early adopters of predictive analytics model. For example, banks use the data collected from credit scores to determine the likelihood of an individual to qualify for a bank loan. The technique has assisted banks to minimize the risks by detecting applicants likely to default the bank loans.

Apart from the bank sector, several organizations in different sectors also use the predictive model to achieve competitive market advantages. For example, sales and marketing department can use historical sale data collected from a specific geographical region to predict probability of sales in the regions. Using historical sales data, an organizational marketing department can target which region and segment to focus their marketing campaigns. More importantly, organizations can use historical data to optimize between price and demand of any product and determine the best pricing for the product.

Don't use plagiarized sources. Get Your Custom Essay on
Application of Predictive Analytics Essay Paper
Just from $13/Page
Order Essay

Predictive analytics can also assist healthcare sector to achieve a better health outcomes. For example, healthcare sector can use the model to predict likelihood that a patient carrying certain type of symptom is suffering from heart attack. The relationships will assist the healthcare to determine the urgency of the treatment.

The police departments also use predictive analytics to reduce crimes. In 1994, the NYPD (New York Police Department) adopted predictive analytical technique to solve crimes. The NYPD developed COMPSTAT known as Computer Statistics to detect likely areas where crimes could occur, and the NYDP uses the GIS (Geographic Information Systems) to map out likely locations in the New York City that crimes can occur, map problem areas and identify hotspots. Typically, the NYDP uses the COMPSTAT to collect large volume of historical data using mathematicians to develop algorithms running against historical data in order to predict future crimes in New York. The strategy is known as predictive policing. Using this strategy, the New York police department is able to reduce automobile thefts, burglaries, and other crimes in the New York City.

Objective of this paper is to explore the application of predictive analytics, its benefits and shortcomings.

Applications, Benefits and Shortcomings of Predictive Analytics

Predictive analysts use the statistical modeling to understand internal and external value of an organization. The model assists organizations to identify patterns and trends and help decision makers to make effective decisions. Typically, predictive analytics is a proven driven force in business intelligence, which assists organizations to enhance competitive market advantages. Application of predictive analytics cuts across different industries ranging from the financial sector to the retail industry. Consumer behavior forecasting is one of the important aspects where predictive analytics is very important. Forecasting is to accurately predict what will happen in the future, however, forecasting is not always based on knowledge and experience, and accurate forecasting is based on the analysis of historical data to extract useful information from data. Predictive analytics is a power tool that organizations employ to do forecasting. Typically, predictive analytics combines powerful analysis technologies with automated discovery to prepare for future based on the analysis of historical data. The decision makers use a large quantity of structured and unstructured data originated from sources such as Customer feedbacks, Call-Center, Websites, Email, and other sources. These data combined together are analyzed to discover threats, and patterns that assist organizations to make decisions on the direction to take. Predictive analytics is based on algorithms, pattern generation, trend analysis and artificial intelligence to enhance future predictions.

Firms can refine prediction using past data to understand consumer behaviors. Using different range of variables, organizations can analyze the demand to arrive at future demand outcomes. For example, Wal-Mart has been able to predict the demand of snowblowers in winter using hard historical data collected from customer demand. Before wide adoption of predictive analytics, the New York Police was the first organization that uses a large-scale predictive model to combat crime in the New York. The authorities were demanding the NYPD to reduce crime in the New York City despite the budgetary restraints. Ability to use limited resources became a high priority for the police department, and thus, the NYPD developed COMPSTAT software, which was an advanced analytical tool to track crime offenders. While data collection and reporting are very critical, however, the tools are not sufficient to enhance public safety. Predictive policing assists the NYPD to effective use scarce resources to achieve better outcomes. One of the benefits of CompStat technology is that the New York police department is able to identify the crime hotspots, which assists the New York police force to quickly respond to locations where crimes are happening. The New York predictive policing model gains a wide acceptance because the NYPD has been able to reduce crimes. (William and Sean, 2008).The CompStat was adopted in 1994 and before the adoption of CompStat, Robbery rates in New York city were 85,892 in 1993. However, in 1998, robbery rates dropped to 39,003. By 2013, the robbery rates in New York city drooped to 19,128. Similarly, murder rates dropped from 1,927 in 1993 to 335 in 2013. Burglary reduced from 100,936 in 1993 to 17,429 in 2013. Typically, the total crime rates in the New York city dropped from 430,460 in 1993 to 7.400 in 2013. The data in Table 1 shows that predictive policing achieves the measurable results.

Table 1: Historical Crime Rate in New York City

















Fel. Assault










Gr. Larceny















Source: NYPD CompStat Unit (2014).

Following the success that NYPD department derives from the CompStat technology, many organizations also use the predictive analytics to enhance market competitive advantages. Evidenced from CompStat model reveals that collecting data intelligently can assist organizations with limited resources to increase revenues. In the last few decades, businesses have also adopted predictive analytical techniques executing their predictive policing model. For example, Marketing and E-commerce companies have learnt to use advance analytical tool to support business intelligence in order to predict, anticipate, and effectively emerge patterns, trends and consumer behavior.

Predictive analytics is the analysis and systematic review of data using automated methods. By using advanced statistical method in combination with exploratory graphics, artificial intelligence, and machine learning tools, an organization can extract useful information from data. By probing information further, it will be possible to accept or reject hypotheses. Moreover, predictive analytics assists organizations to predict trends, relationships, sequence, and patterns on the data, which can be used to anticipate actions.

Over the years, a company such as Wal-Mart has understood the importance of predicting future demand. For example, Wal-Mart has been able to shift its supply chain by sending bottle water, duct tape and Pop-Tarts to the affected locations during the storm. Essentially, products such as bottled water and duct tape make sense for emergency preparations and responses. While the Pop-Tarts product seems very odd, however, Wal-Mart has used predictive analysis to increase market advantages of Pop-Tarts after analyzing the historical data. Wal-Mart is able to match weather events with the Pop-Tarts. Analysis of the past data assists Wal-Mart to increase demand for the product by accurately forecasting likely of the storm to occur in a particular location. With predictive data analytics, Wal-Mart has been able to adjust its supply chain and move the strawberry Pop-Tarts to locations where there will be a large or bad weather. Thus, accuracy of Wal-Mart speculation increases consumer demand because the company has already moved the products to locations where there is a likely occurrence of storm.

Moreover, insurance companies use predictive analytics to detect fraud and predict the pattern of frauds using a statistical model for fraud prevention. Moreover, auto insurance company can use predictive analytics to determine an accurate premium to charge customer. Underwriters also use predictive analytics to predict bankruptcy, default, and a chance of illness of applicants.

Kent, (2006) argues, “Predictive analytics consists of two major components – advanced analytics and decision optimization. Decision optimization and advanced analytics use comprehensive portfolio of sophisticated statistical techniques and data mining algorithms.” (p 40). According to Kent, predictive analytics also uses different tools to predict future trends, and link analysis algorithms, which is the process of developing network of cases and relationships to build patterns and trends. Moreover, regression analysis assists in predicting a value-using dependent and independent variables, estimating forecasting to build relationships between variables. On the other hand, Maciejewski, et al. (2011) point out that predictive analytics can assist analysts to predict future hotspots using statistical analytical view to facilitate forecasting. In the present business environment, analysts are searching for unexpected events to trigger alerts. The strategy is to drill down data to redistribute resources, control problems and confirm alert. The authors argue that predictive analytics paradigm is an effective tool to achieve these goals.

Maciejewski, et al. (2011) demonstrate effective impact of predictive analytical tool by collecting historical data “to detect public health emergencies before such an event is confirmed by diagnosis or overt activity.” (P 2). For example, collecting historical data from the listed organizations assists in managing public health emergency:

ISDH (Indiana State Department of Health), and PHESS “(Publi Health Emergency Surveillance System).” (Maciejewski, et al. 2011 P. 2).

The authors have able to locate the hotspots in the United States where certain category of diseases is likely to occur, which could assist hospitals to make future planning and respond to patients’ treatment in the locations. Budale and Mane (2013) also point out that banks are using predictive analytics to improve their customer relationships. In the contemporary financial environments, banks are facing intense competition due to an increase in customer demand. Typically, customers can switch to another bank if their banks do not deliver the service required. Thus, predictive analytics is the strategy banks employ to retain their customers.

Budale, & Mane, (2013) argue that several banks use predictive analytics to mine financial and operational data and utilize them for their business benefits. In the present technological advancement, customers are using high-end technology for their transactions, and there is a little interaction between customer and bank staff. Thus, banks are required to anticipate customer requirements in order to provide them with better services since “customer acquisition is more costly than customer retention.” (Budale, & Mane, 2013 p 508). Thus, banks are required to use predictive analytics to predict the customer needs. Essentially, predictive analytics assists banks to make faster and better decision using automated business process. Moreover, analytics assists banks to make faster decision.

“The predictive analytics also helps banks grow customer base and retain most profitable customer, continuously improve operational efficiency, prevent fraud and manage risk and compliance efforts, transform and automate financial processes.” (Budale, & Mane, 2013 p 508).

Apart from assisting organization to make important prediction about the business process, predictive analytics also assists organizations to reduce churn and attrition. Typically, predictive analytics assists in enhancing relationships between profile information and data to identify a propensity of a customer churning out. Moreover, the model can assist organizations to identify customer lifetime value. The strategy will assist organizations to gain customers’ loyalty and maximize customers’ values.

Delinquency management is another area where predictive analysis is very beneficial, and the model is an appropriate strategy for credit recovery. Moreover, predictive analytics can assist organizations to make a cross selling of their products. Cross selling assists banks to reduce marketing costs by promoting product and services to genuine customers who are interested in the product. More importantly, predictive analytics is beneficial in determining churn probabilities, customer lifetime value and product propensity. Churn probability determines method to calculate the loss of customers to other company, and product propensity is the likelihood of clients buying product offered.

The “customer Life Time Value (CLTV) is the customer’s potential monetary worth through course of their relationship with business, calculated across entire CRM life cycle, including all functions, processes, channels, roles, campaigns, and touch points.” (Budale, & Mane, 2013 p 508).

Despite the benefits that organizations can derive from predictive analytics, however, predictive analytics is very complex to implement. The specialized mathematical and statistical knowledge is required to develop a model and predict the future outcomes. Moreover, a considerable knowledge of mathematical and statistical knowledge is required to implement analytical tools and software package required to apply the model. Analysts with high experience in R. Programming, Java, and C++ programming do predictive analytics. The R. knowledge is very critical for predictive analysis because R. is a free software programming used for graphics and statistical computing. Typically, statisticians and data miners widely use R. programming to carry out data mining techniques to develop statistical software. Moreover, R delivers a wide variety of graphical and statistical techniques, which include linear and nonlinear modeling to carry out statistical tests, classification, clustering, and time-series analysis. Another important aspect of R-programming is that it is an open source and very easy to use. The use of R. programming is similar to the application of Java or C++ programming language because R. has in-built functions for modeling and calculations. Essentially, algorithms are very critical for predictive analytics and R. programming is very useful to create an algorithm. Presently, R has the ability to create algorithms and has approximately 400 algorithms within the R. library ranging from linear regression to advanced model. Typically, there is no limit in the application of R- programming language. One of the usefulness of R. programming is its application for predictive modeling, which assist analysts to calculate product propensity, and churn probability.

Apart using R. programming for analysis, predictive analysts also use regression technique to carry out the analysis. Essentially, regression model is one of the mainstays of predictive analytics using the mathematical equation to establish interaction between different variables. Depending on the type of application, a predictive analyst can use a wide variety of regression model. For example, linear regression model establishes relationships between dependent and independent variables to make a reasonable prediction. Using OLS (ordinary least square, it is possible to check a statistical significant of different variables.

Time series model is another analytical tool that analysts use for the predictive analytics. The time series analysis uses historical data to forecast future behaviors. The model bases its assumption that the behaviors of data over time will have an autocorrelation for future behaviors. Typically, time series uses moving average and autoregressive model for forecasting.

Other statistical tools that predictive analysts use to predict future events are:

Discrete choice model,

Logistic regression

Multinomial Logistic regression,

Probit Regression.

Time Series Model

Survival or duration analysis.

Based on the sophistical techniques in carrying predictive analytics, organizations that intend to use predictive analytics will need to hire an expert or use an external vendor to carry out the analysis.

Intruding in consumer privacy is another shortcoming on the use of predictive analytical tool. Essentially, the predictive analytics uses data mining to collect historical data and data mining process can lead to the intrusion of customer’s confidential data. For example, analysts can come across customers’ confidential data that include Social Security Number (SSN), bank account number, credit card information and other sensitive information, which can lead to a private data breach. The popularity of predictive analytics in the past few years has made many customers to reluctantly submit their data during online shopping. Another issue is that the government can get access to consumer data in the course of the investigation of a consumer thereby land on the private data on other consumers.

Despite the shortcomings identified in the use of predictive analytics, predictive analytics is still an effective analytical tool that organizations employ to gain competitive advantages. Organizations use predictive analytics to develop an analytical customer relationship management. For example, organizations can use historical data to predict sales, design marketing campaign, and customer relationship services. Moreover, predictive analytics assists medical professionals to achieve health outcomes with clinical decision support. For example, health professionals can make effective medical decisions using predictive analytics to improve healthcare system.


This paper investigates the application of predictive analytics, its benefits and shortcomings. Predictive analytics is a statistical technique of using historical data to predict future outcomes. Typically, predictive analytics is based on the assumption that historical data can be used to forecast future events. The paper shows that the NYDP is the first organization in the United States that uses predictive analytical tool to reduce crimes in the New York City. More important, banks also use the predictive analytics to determine the credit scores of loan applicants. The strategy is to determine whether a loan applicant has ability to repay the loan. The successes derived by banks and NYPD from the use of predictive analytics have made many organizations to use predictive analytical tool to achieve competitive market advantages. For example, organizations use past sales data to predict on future sales forecasts. Despite the benefits derived from the application of predictive analytical tool, skilled mathematicians and statisticians are required for the effective applications of predictive analytics. Typically, the knowledge of R. programming, regression model and other statistical tools are required for application of the model. Thus, organizations intending to use predictive analytical tool for forecasting are required to train in-house staffs to carry out the tasks. However, external vendors could also be used to carry out the tasks in order to cut costs. However, the use of external vendors may not be advisable in a work where confidential customer data are involved.


Budale, D. & Mane, D. (2013). Predictive Analytics in Retail Banking. International Journal of Engineering and Advanced Technology (IJEAT), 2 (5): 508-509.

Kent, B. (2006). Predictive Analytics: Algorithm Nirvana. DM Review,16(30):40.

Maciejewski, R. Hafen, R. Rudolph, S. et al. (2011).Forecasting Hotspots — A Predictive Analytics Approach. IEEE Computer Society.Issue 10.

NYPD CompStat Unit (2014). CompStat. Police Department City of New York. 21(22).

William J.B. And Sean W.M. (2008). Police Performance Management in Practice: Taking COMPSTAT to the Next Level, Policing, 2( 3): 259 — 265.