business IT project 1I have 2 Business IT case study assignments, i’m posting one now to understand the work and then I will send the other one . I will post the questions and upload the case studies and the materials, let me know if anything is unclear or if you have any questions.Answer the questions below in the context of each of the following five cases (all pdf files) of real-world business information systems.QUESTIONS
Sport AnalyticsUsing Open Source Logistic Regression Software to ClassifyUpcoming Play Type in the NFLRobert E. BakerTed KwartlerAbstractThe purpose of this study was to utilize data analytics as means to classify NationalFootball League offensive play types. The open source software R was employed tocreate a logistic regression based on data for the Cleveland Browns and PittsburghSteelers from 13 recent seasons. The regression is based on all first, second, andthird downs within regulation play, totaling 26,310 data points. The initial algorithms classify rush or pass for each offense. Revealed through differing coefficients of the independent variables, each team shows a slightly different approachto play selections in response to in-game situations. Identifying the driving factorsto play selection is possible by isolating each attribute within the regression. Further examination could yield improved precision to control for changes in headcoach, offensive coordinators, player personnel and other factors such as weatherbecause these may influence play type. Logistic regression shows promise as anin-game aid to determining opponent behavior. Specifically, Cleveland’s offensiveplay selection algorithm was correct for 66.4% of plays versus 66.9% for Pittsburgh. Use of open source software and logistic regression of NFL play selectioncould be beneficial in aiding future game decisions. Further research is recommended to explore possible improvement of the algorithm accuracy.Keywords: sport analytics; sport management; data mining; NFL; regression43Robert E. Baker is an associate professor in the Department of Sport and Recreation Studiesat George Mason University.Ted Kwartler is Director of Customer Success at DataRobot.Please send correspondence to Robert Baker, [email protected] of Applied Sport Management Vol. 7, No. 2, Summer 2015Sport Analytics44IntroductionSport is big business. This top-10 segment of the global economy is estimatedat $440 to $470 billion in North America alone (Fry & Ohlmann, 2012; PlunkettResearch, 2014). Reflecting this economic position, sport organizations parallelbusiness processes in any major economic segment. Decisions in sport organizations are increasingly informed by, and derived from data analysis (Andrew, Pedersen, & McEvoy, 2011). The use of statistical analysis is an essential component indata driven decision making. Sport is a rapidly growing arena for the applicationof analytics (Fry & Ohlmann, 2012). The essence of sport analytics includes managing data, using predictive analytics, and informing decision makers to providea competitive advantage (Alamar, 2013). This reveals a continual analysis processin sport settings wherein situation-specific information informs analytical algorithms, which in turn guide data collection and analysis. Once analyzed, leadersemploy data in decision making, which yields results that feed back into situational analyses (see Figure 1).Moneyball (Lewis, 2004) yielded perhaps the most visible example of the application of sport analytics. It is the story of the management strategy of BillyBeane, of the Oakland A’s, who employed statistically driven, evidence-basedpractice to enhance efficiency in winning baseball games (Cullena, Myera, & Latessaa, 2009; Surendra & Denton, 2009). The book was a New York Times (2011)bestseller, and was developed into a successful movie grossing over $140 million(Nash Information Services, LLC, n.d.). While the Moneyball story obviously hadentertainment value, the story of the evolution of sport analytics has become valued in many sport settings. The ability to generate and apply statistical interpretations to enhance efficiency is central to the use of data in sports (Hakes & Sauer,2006). Beyond this application of Sabermetrics, the mathematical and statisticalFigure 1..Continuous Analysis in SportSitua’on-‐specificplay
parameters
informs
algorithm
Data
collec’onand
sta’s’cal
analysis
Applica’onof
analysis
through
decision
making
Resultant
performance
outcomes-‐
successor
failure
Figure 1. Continuous Analysis in SportBaker and Kwartler45analysis used in Major League Baseball (MLB), examples of the use of data analytics in sport abound (Costa, Huber, & Saccoman, 2008; Davenport & Harris, 2007).This developing phenomenon found its way to the English Premier League (EPL),the National Basketball Association (NBA), the Professional Golfers’ Association(PGA), the National Hockey League (NHL), and the Rugby Union, among manyother sport organizations (Alamar, 2013; Brugha, Freeman, & Treanor, 2013; Chu,2013; Fry & Ohlmann, 2012; Gerrard, 2007; Goddard, 2005). Data analytics areused in a variety of capacities in sport organizations. Analytics can enhance productivity and efficiency in sport, whether evaluating prospective talent, individualizing customer service, or informing managerial and coaching decisions.“Big Data,” Data Mining, and Spot AnalyticsBusinesses, governments, communities, and individuals access huge collections of data, trillions of bytes known as “Big Data,” which capture valuable information and provide them such benefits as improved efficiency and effectiveness,ultimately leading to a competitive advantage (Bryant, Katz, & Lazowska, 2008).To pursue this advantage, novel applications such as individualized customizationwill result from the capture and analysis of data in order to meet the demands toharness “Big Data” (Borkar, Carey, & Li, 2012; Brown, Chui, & Manyika, 2011).The rise of the phenomenon of “Big Data” may result in better tools, goods, andservices; yet, it may yield invasions of individual privacy. Still, “Big Data” mayhelp enhance our understanding of trends, communities, and individuals (Boyd& Crawford, 2012).There is no specific size that determines how big “Big Data” has to be (Manyikaet al., 2011). It varies by economic sector and intended purposes. Only by fuelinginnovation, competition, and productivity will “Big Data” manifest its value. As“Big Data” becomes more available and more usable in informing strategic decisions, the mining of relevant data is also being used to guide decisions in sport(Lohr, 2012).Data mining involves the quest for and discovery of valuable structures inlarge datasets (Hand, 2007). While a global concern of data mining is in “Big Data,”another aspect is in small-scale structures that can inform local decisions. Whilesport analytics has a broad concern with “Big Data,” it is in the local arena whereit can be essential to sport managers (Hand, 2007). For example, mining data tosupport the management of customer relationships can result in enhanced customer loyalty, specifically targeted prospects, and newly identified markets (Berry& Linoff, 1999). The array of techniques for data mining can be used to producebetter decisions across many areas of any sport organization, from marketing andcustomer support strategies to player recruitment (Berry & Linoff, 1997). Basically, data mining can be used to achieve strategic results by formulating the problem, analyzing the data, interpreting the results, and utilizing this information(Berry & Linoff, 1999).Sport Analytics46The discovery of new knowledge yields an increased pool of data to mine,and the mining of data yields new knowledge. The amount of accessible data hasdramatically escalated through enhanced ability for data generation and collection (Han, Kamber, & Pei, 2006). The growth in the generation of data is dueto more computerized transactions and increased use of digital cameras and barcodes. The collection of data is intensified by scanned text and images, satellitesystems, and the vast amounts of data on the web. This growth in available datamandates that researchers and practitioners find ways to transform it into usefulinformation, and ultimately competitive advantage and improved performance(Han, Kamber, & Pei, 2006). Therefore, the demand for intelligent tools that automatically assist in transforming data into useful knowledge is intense (Fayyad,Piatetsky-Shapiro, Smyth, & Uthurusamy, 1996). There is an array of analyticaltools for effective model building (Guazzelli, Lin, & Jena, 2012). One such dataanalysis tool is the open source interactive statistical software R. R is an opensourced computing and graphics platform that comprises an integrated suite ofsoftware for “data manipulation, calculation, and graphical display” (R Foundation, 2014). Developed by the R Foundation, a nonprofit organization workingin the public interest, R is available as free software that runs on a wide variety ofUNIX platforms, Windows, and MacIntosh operating systems.Data mining has a direct relation to statistical analysis (Hand, 2007). Aspectsof data mining, such as classifying, clustering, trend and deviation analyses, anddependency modeling, intersect with the realm of statistics (Fayyad et al., 1996).Business intelligence (BI), the term used to describe analytic applications, essentially involves data input and output (Watson, & Wixom, 2007). While both inputand output are important, the focus of many organizations is in the application ofdata output to be used in decision-making processes. This focus has led to increasing interest in data mining and the application of statistical algorithms (Guazzelli,Stathatos. & Zeller, 2009).Data mining has created a need for statistical algorithms and open sourcesoftware solutions that have prompted predictive analytics to become a standardapproach to BI (Guazzelli et al., 2012). A challenge associated with data miningand analysis is the explosive increases in the volume of data (Guazzelli et al., 2012).If properly utilized, these knowledge management tools can inform strategic andtactical decision making despite the abundance of both relevant and irrelevant information. Creating competitive advantage in many arenas as knowledge discovery tools, data mining and predictive analytics allow sport managers to utilize theinput to yield actionable outputs (McCue, 2005). Decision management systems,as they are called, utilize these knowledge management tools to yield effective decisions that produce competitive advantage (Taylor, 2012).Pursuing competitive advantage is a fundamental principle in the realm ofsport. The uses for data mining and subsequent predictive analytics in sport whileevident, remains incomplete. There are many sport settings, both in the front of-Baker and Kwartler47fice and on the field, where more efficient and effective decision-making systems,powered by data mining and predictive analyses, may enhance productivity. Thisstudy explores an on-field application of these techniques within the NationalFootball League (NFL).PurposeThe purpose of this study was to utilize data analytics as means to classifyNFL offensive play types. This research more specifically examines the applicationof logistical regression analysis utilizing open source data in comparing two NFLteams’ play selections between the 2000 and 2012 seasons.Specifically, the Cleveland Browns and the Pittsburgh Steelers, the longestrunning rivalry in the American Football Conference (AFC), were selected forreview. In this timeframe, the Browns have perennially been an unsuccessful teamas measured by wins (71). In contrast, the Steelers have been extremely successful,not only winning 64 (135-71) more games than the Browns but also the 2006 and2009 Super Bowl championships. However, in some respects, these teams, separated by 135 miles, are similar. Both teams play in the same conference, therebysharing many opponents, have similar “blue-collar” work cultures, and representsimilar-sized markets. This study had as an additional intent to compare a “bad”team to a “good” team while attempting to negate other factors. The primary objective is to demonstrate logistic regression as a viable methodology to help NFLdefensive coordinators assess the probability of a rush or pass by an opposing offense.MethodologyData CollectionData was acquired from www.armchairanalysis.com, and primarily camefrom the “CORE” data file. This file contains over 500,000 individual records.Each record represents a single NFL play and collected attributes. The data wasreduced and segmented in this research using the following parameters. Fourthdown plays were removed because these are overwhelmingly punts and field goals,or otherwise extremely unique game situations. Play types such as kickoffs wereremoved. This resulted in a binary relationship, “rush” or “pass,” being kept foranalysis. The overtime was removed by excluding any number larger than four as aquarter. Overtimes were removed because they represent a unique game situation.The resulting data set was then partitioned according to offense. This resultedin 12,187 Cleveland (CLE) play records and 14,123 Pittsburgh (PIT) play records.The models were built individually from each partition. Each record had 10 attributes that were selected for model building, along with a variable that changedplay type to binaries: 0 (RUSH) or 1 (PASS). Table 1 shows the attribute name anddefinition. Table 2 shows example records with both used and unused attributesfrom the “CORE” data set.Sport Analytics48The data was examined using a logistical regression analysis so as to providethe logistical odds of a binary event and the resulting probability. Logistic regression differs from linear regression in that it classifies the dependent variable ina binary relationship. A linear regression predicts continuous outcomes such asgames scores. Instead of a game score, it classifies winning or losing. The generallayout of the proposed logistic regression is as follows:F(log odds) = intercept + coefficient(attribute 1) + coefficient(attribute2)…+coefficient *(attribute n)+ e, where “n” is the total number of attributes and“e” is the error term.AnalysisThe open source R software was use for the logistical regression analysis. TheR statistical software, freely available at http://www.r-project.org/, is an interactiveprogramming language with extensive capabilities for quantitative analysis (Maindonald & Braun, 2009). The R scripts can be obtained by contacting the author.The R software was used to read the modified data set. Then, R created the coefficients for each attribute by fitting a logarithmic curve to the records. A visualcomparison of the differences between linear and logistic regression is illustratedin Figures 2 and 3.GameID Season PlayID Off Def Type Driveseq Qtr Min Off _pts DefptsOff_TODef_TODWN YTG YFOG12 2000 1795 PIT BAL 1 1 1 15 0 0 3 3 1 10 3812 2000 1796 PIT BAL 0 2 1 15 0 0 3 3 2 6 42Table 1.1 Example RecordsTable 2Example RecordsAttribute Name Stands for Definitiondrive_seq Drive Sequence The number of the play in the drive sequence.QTR Quarter The play takes place in 1 of 4 quarters of the regulargame time so this is a classification of 1 to 4.MIN Minute The exact minute of the play start within the quarter.Off_pts Offensive points The points of the offensive team at the beginning of theplay.Def_pts Defensive points The points of the defensive team at the beginning ofthe play.Off _TO Offensive time outs The remaining time outs of the offensive team.Def_TO Defensive time outs The remaining time outs of the defensive team.DWN Down The down of the play, due to prior data reduction this islimited to 1,2 or 3.YTG Yards to Go The yards to go in order to get a first down or reach thegoal line from the line of scrimmage.YFOG Yards from owngoalThe distance between the line of scrimmage and theoffensive team goal line.Table 1.0 Attribute Dictionary used in the ModelsTable 1Attribute Dictionary Used in the ModelsBaker and Kwartler49Figure 2. Figure 1.0LogarithmicOdds Logarithmic OddsThe logistic regression calculates the logistical odds (log odds) of an event occurring. However, in practical terms, this is not very helpful. Thus, once log oddshave been calculated, it is used as input to get a more practical outcome, knownas probability. Probability is the percentage likelihood of an event occurring ornot occurring. The range of probability approaches 0% to 100%. The equation tomove from log odds to probability is:(e^log odds)/(1+(e^log odds), where e is the natural log or ~2.718.When plotting the probability and log odds as a scatter plot, a curve from 0 to1 occurs. This is illustrated in Figure 3. An event occurs somewhere on the probability curve and is then classified as a rush or pass based on a cutoff probabilityparameter. Commonly, the probability cutoff is .5 or 50%.Figure 3. Figure1.1Fictitious Linear Regression : FictitiousLinearRegressionSport Analytics50Resulting ModelUsing R, a coefficient table can quickly be calculated among the ~26,000 records. Table 3 shows the CLE and PIT logistic regressions based on the data set.Although this model calculates the log odds, which is of limited value, a comparison can still be made. All coefficient signs are the same with the exception of thequarter (QTR). This means that as the quarter increases the CLE model will showa higher log odds (and resulting probability) of passing in the model. Converselythe PIT model shows a diminishing desire to throw as the game quarters increasedue to the negative sign. Further research could explore if this phenomenon is theresult of the Browns trailing more often than the Steelers late in games. Howeverthis is out of scope of this research.Model EvaluationClassification models are often compared using a confusion matrix and applied to a holdout validation data set (see Tables 4 and 5). Since we have generatedtwo models, two separate matrices are needed to compare the model classifications to the actual outcomes in the data set. Typically, classification models arepartitioned a priori into training and validation sets. In this case, all records wereused for the supervised learning of the model. While not as scientific, the additional records likely improve model development. The next truthful validationset could be obtained as the 2013 NFL season progresses. In the meantime, theconfusion matrices are directionally sound to assess accuracy.The Cleveland Browns the model was 66.4% accurate. The model classified8,090 play types correctly and 4,097 incorrectly. The PIT model was 66.9% accurate. The model classified 9,442 play types correctly and 4,681 incorrectly.Figure1.2 Probability CurveFigure 4. Probability CurveBaker and Kwartler51Attribute Name CLE PITIntercept -2.073 -2.008drive_seq 0.025 0.017QTR 0.065 -0.128MIN -0.009 -0.011Off_pts -0.046 -0.043Def_pts 0.046 0.055Off _TO -0.261 -0.333Def_TO 0.094 0.183DWN 0.868 0.980YTG 0.114 0.142YFOG -0.001 0.003Table 1.2 Logistical Regression CoefficientsTable 3Logistical Regression CoefficientsCLE MODEL Actual Play was Pass Actual Play was RushModel ClassifiedPass5,106 2,289Model ClassifiedRush1,808 2,984TABLE 1.3 CLE Confusion MatrixTable 4CLE Confusion MatrixPIT MODEL Actual Play was Pass Actual Play was RushModel ClassifiedPass4,765 2,140Model ClassifiedRush2,541 4,677TABLE 1.4 PIT Confusion MatrixTable 5PIT Confusion MatrixSport Analytics52AttributeNamePIT PlayID3447 Log ODDSIntercept -2.008 -2.008drive_seq 0.017 1 0.017QTR -0.128 4 -0.512MIN -0.011 14 -0.154Off_pts -0.043 10 -0.430Def_pts 0.055 13 0.715Off _TO -0.333 1 -0.333Def_TO 0.183 3 .549DWN 0.980 1 0.980YTG 0.142 10 1.42YFOG 0.003 30 0.09TOTAL 0.334TABLE 1.5 Example PIT Play and Log Odds CalculationTable 6Example PIT Play and Log Odds CalculationExample CalculationsIn this section, a play for each team is examined and a third play is used tocompare teams side by side. Pittsburgh’s play identified as ID 3447 was againstBaltimore Ravens’ defense and resulted in a pass. It was the first play of the drivein the 14th minute of the fourth quarter. Pittsburgh was trailing Baltimore by ascore of 10 to 13. Additionally, the Steelers only had a single timeout left versusthe Ravens’ three. Specifically, it was first-and-10 on the Pittsburgh 30-yard line.The log odds calculate to .334 in this play scenario using the operators outlined in the Methodology section. To get to probability, one must still take the stepusing the natural log as outlined in the Methodology section:(2.718^.334)/ (1+(2.718^.334)) = .582 or 58.2% probability of a PASS.Cleveland’s play identified as ID 531127 was against Cincinnati’s defense andresulted in a rush. It was the second play of the drive in the sixth minute of thefirst quarter. Cleveland was losing to Cincinnati by a score of 0 to 7. Both teamshad three timeouts. Also, it was second-and-3 on the Cleveland 23-yard line.Baker and Kwartler53AttributeNameCLE PlayID 531227 Log ODDSIntercept -2.073 -2.073drive_seq 0.025 2 0.05QTR 0.065 1 0.065MIN -0.009 6 -0.054Off_pts -0.046 0 0Def_pts 0.046 7 0.322Off _TO -0.261 3 -0.783Def_TO 0.094 3 0.282DWN 0.868 2 1.736YTG 0.114 3 0.342YFOG -0.001 23 -0.023TOTAL -0.136TABLE 1.6 Example CLE Play and Log Odds CalculationTable 7Example CLE Play and Log Odds CalculationThe log odds calculate to -0.136 in this play scenario. The next step is to usethe log odds to create the probability of the Pass:(2.718^-0.136) / (1+ (2.718^-0.136)) = .466 or 46.6% probability of a PASS.A third example serves as a direct comparison between the teams. This is areal play from the Pittsburgh data set and is similar to plays in the Cleveland oneas well. In this situation, and ones like it, the gap between the teams can be large.In this case, the result is a classification of a Pittsburgh Pass versus a ClevelandRush.Using the log odds for each team, PIT has a 50.1% probability of a pass whileCleveland had a 38.8%.Discussion and ImplicationsThe results of this study demonstrate that open source software can be effectively used in the analysis of situation-specific play selection in the NFL. Itprovides evidence that logistic regression has promise as a classification systemfor opposing NFL offenses. The analysis of large amounts of data regarding playselection can inform decisions by head coaches, defensive coordinators, positioncoaches, opposing coaches, and even player personnel managers. Head coaches,defensive coordinators, and position coaches can better develop a game plan andprepare athletes for opponents’ anticipated play selection. Conversely, offensivecoaches might choose to alter their documented play selection pattern simply bybeing informed of it. Player personnel managers may identify and pursue talentspecifically based on an analysis of situation-specific play selection, both by theirown team and by their opponents.Sport Analytics54This research demonstrated that data analytics can add value in the decisionmaking process in sport settings such as the NFL. It also demonstrated that humans are central to the effective implementation of sport analytics. Slaton (2013)noted that “…the entire sports organization, from the lowliest assistant coach andmarketing employee to the most senior leader needs to adopt the analytics philosophy if it is to be truly effective” (p. 1). Coaches and sport managers are centralin the collection of data, in the analysis of that data, and in the application of thatdata analysis.Quantitative analysis has been shown to yield valuable information that sportmanagers and coaches can use to inform their decisions (Borrie, Jonsson, & Magnusson, 2002). There is an abundance of data available to support both on-thefield and in-the-front-office decisions in sport. It is incumbent on sport managersto obtain, analyze, and utilize appropriate data within their organization’s knowledge management systems.In the coming years, the United States will need up to 190,000 more expertanalysts and an additional 1.5 million managers, including sport managers, whoare prepared to use data to inform decisions (Manyika et al, 2011). The applicationof analytics in sport is no longer the niche pursuit of a few visionaries such as theOakland A’s Billy Beane, the Houston Rockets’ Daryl Morey, or the Kraft Group’sAttributeNamePIT CLE ComparisonPlayPITT LogODDSCLE LogODDSIntercept –
2.008
2.073 -2.008-2.073drive_seq 0.017 0.025 1 0.017 0.025QTR –0.1280.065 1-0.1280.065MIN –
0.011
0.0099-0.099-0.81Off_pts –
0.043
0.046000Def_pts 0.055 0.046 3 0.165 0.138Off _TO –
0.333
0.2613-0.999-0.783Def_TO 0.183 0.094 3 0.549 0.282DWN 0.980 0.868 1 0.98 0.868YTG 0.142 0.114 10 1.42 1.14YFOG 0.003 –0.001360.108-0.036TOTAL 0.005 -0.455TABLE 1.7Table 8Direct Team ComparisonBaker and Kwartler55Jessica Gelman, but rather it is a mainstream practice that sport organizationshave embraced (Slaton, 2013). For example, over 2,700 interested researchers,analysts, and sport managers annually attend the Sloan Conference on the latestdevelopments in sport analytics (Slaton, 2013).As evidence of the successful application of sport analytics builds, a new styleof sport manager is evolving in sport organizations (Fry & Ohlmann, 2012). Thesenew sport managers and coaches are proficient in the application of data analytics. The organizations they work for are stimulating the expansion of data miningand predictive analysis in sport settings. The ongoing development of analytics insport settings reflects the trend in other market segments and is inevitable (Alamar, 2013). The proper utilization of data management, predictive analyses, anddata-driven decision-making by sport managers occur in pursuit of a competitive advantage (Alamar, 2013). Using the data analyzed in this study, coaches andmanagers seeking a competitive advantage can employ data-generated knowledgeabout opponent tendencies in their decisions on player personnel, formations,and play calls in these specific game situations.This study supports the contention that the analysis of data in sport can bemost useful in its application by qualified sport leaders (Lohr, 2012). In that context, the increasing effectiveness and acceptance of sport analytics, while informing decisions, does not negate the value of human insights and actions (McAfee &Brynjolfsson, 2012). Sport managers as decision makers, informed by data analytics’ knowledge management tools, remain essential. They are crucial in collectingappropriate data and in analyzing that data accordingly. More so, effective coachesand sport managers of the future must be adept in the successful strategic application of the analysis to specific situations. For example, in this study, it remains incumbent on head coaches, offensive and defensive coordinators, position coaches,and player personnel managers to determine the best way to effectively employ thedata analyses in the decision-making process. Additionally, their intended use ofthe data can inform the selection and analysis of algorithmic factors,The process of continual analysis in sport is the cyclical progression throughwhich analytics are effectively utilized in sport settings (see Figure 1). Initial situation-specific parameters (e. g. plays executed) are identified and inform the development of the analytical algorithm. Within the identified parameters, data arecollected and statistically analyzed. Sport leaders (e. g. coaches) apply the analysisthrough the process of informed decision-making in similar situations. The resulting performance outcomes (i. e. success or failure) in like situations inform thepertinence of the identified parameters and influence the development or maintenance of the employed algorithms.This study provided evidence of situation-specific tendencies for two teams inthe NFL. The study was delimited to only the Cleveland Browns and PittsburghSteelers. All other teams were excluded, as were other professional sport leagues.Clearly, this analytical approach could be easily applied to any team, in any sport,Sport Analytics56within any situation-specific parameters. In explicit game situations, such as timeremaining or score, first-down plays, plays in the “red zone,” special team playscould be similarly analyzed. Overtime plays were not analyzed, but could be analyzed to inform decisions on that specific situation. Likewise, fourth-down playswere excluded, but could be reviewed in a separate analysis.To better inform the application needs of coaches and sport managers, a researcher could improve the utility and/or accuracy of the model by incorporatingnew attributes, thereby changing the data set. Changes in coaching staff, playerpersonnel, and even opponents’ style of play occur regularly in the NFL. Analyses incorporating algorithms that address these factors can be developed. For example, when Tim Tebow replaced Kyle Orton as the Denver Broncos quarterback,the number of rushes increased significantly. Judiciously excluding these plays asoutlier situations may have a boosting effect. In addition, games with inclementweather may often have more rushes. So, adding an attribute for weather conditions may improve the analyses’ utility. Additionally, a researcher could changethe data structure to improve effectiveness. Coaches, offensive coordinators, andaudible-calling quarterbacks, such as Peyton Manning, likely behave consistentlyin similar game situations despite the team. Thus, partitioning the data from teambased to a specific decision maker (e.g., quarterback) may further improve thepredictive accuracy.ReferencesAlamar, B. C. (2013). Sport analytics: A guide for coaches, managers, and other decision makers. New York, NY: Columbia University Press.Andrew, D., Pedersen, P., & McEvoy, C. (2011). Research methods and design insport management. Champaign, IL: Human Kinetics.Berry, M., & Linoff, G. (1999). Mastering data mining: The art and science of customer relationship management. New York, NY: John Wiley & Sons, Inc.Berry, M. J., & Linoff, G. (1997). Data mining techniques: For marketing, sales, andcustomer support. New York, NY: John Wiley & Sons, Inc.Borkar, V., Carey, M., & Li, C. (2012). Inside “big data management”: Ogres, onions, or parfaits? New York, NY: Proceedings of the 15th International Conference on Extending Database Technology 3-14.Borrie A., Jonsson, G. K., & Magnusson, M. S. (2002). Temporal pattern analysisand its applicability in sport: an explanation and exemplar data. Journal ofSports Sciences, 20(10), 845–852.Boyd, D., & Crawford, K. (2012). Provocations for a cultural, technological, andscholarly phenomenon. Information, Communication, and Society, 15(5),662–679.Brown, B., Chui, M., & Manyika, J. (October, 2011). Are you ready for the era ofbig data? McKinsey Quarterly, pp. 24–35.Baker and Kwartler57Brugha, C. M., Freeman, A., & Treanor, D. (2013). Analytics for enabling strategyin sport. OR55 Annual Conference Keynote Papers and Extended Abstracts,138–150.Bryant, R. E., Katz, R. H., & Lazowska, E. D. (2008). Big Data Computing: Creatingrevolutionary breakthroughs in commerce, science, and society. ComputingResearch Initiatives for the 21st Century, Computing Research Association.Retrieved from http://www.cra.org/ccc/files/docs/init/Big_Data.pdf.Chu, B. (February 25, 2013). Dr. Wayne Winston’s work may do to the NBA whatSabermetrics did to MLB. Yahoo!Sports. Retrieved from http://sports.yahoo.com/news/dr-wayne-winstons-may-nba-sabermetrics-did-mlb-155600962–nba.htmlCosta, G. B., Huber, M. R., & Saccoman, J. T. (2008). Understanding Sabermetrics:An introduction to the science of baseball. Jefferson, NC: McFarland & Company, Inc.Cullena, F. T., Myera, A. J., &Latessaa , E. J. (2009). Eight lessons from Moneyball :The high cost of ignoring evidence-based corrections. Victims and Offenders:An International Journal of Evidence-based Research, Policy, and Practice, 4(2),197–213.Davenport, T. H., & Harris, J. G. (2007). Competing on analytics: The new scienceof winning. Boston, MA: Harvard Business Review Press.Fayyad, U. M., Piatetsky-Shapiro, G., Smyth, P., & Uthurusamy, R. (1996). Advances in knowledge discovery and data mining. Menlo Park, CA: AAAI Press.Fry, M. J., & Ohlmann, J. W. ( 2012 ). Introduction to the special issue on analyticsin sports, part I: General Sports Applications. Interfaces, 42(2), 105–108.Gerrard, B. (2007). Is the Moneyball approach transferable to complex invasionteam sports? International Journal of Sport Finance, 2(4), 214–230.Goddard, J. (2005). Regression models for forecasting goals and match results inassociation football. International Journal of Forecasting, 21, 331–340.Guazzelli, A., Lin, W-C., & Jena, T. (2012). PMML in action: Unleashing the powerof open standards for data mining and predictive analytics. Paramount, CA:CreateSpace.Guazzelli A., Stathatos K., & Zeller, M. (2009). Efficient deployment of predictive analytics through open standards and cloud computing ACM SIGKDDExplorations, 11(1), 32–38.Hakes, J. K., & Sauer, R. D. (2006). An economic evaluation of the Moneyball hypothesis. The Journal of Economic Perspectives, 20(3), 173–185.Han, J., Kamber, M., & Pei, J. (2006). Data mining: Concepts and techniques (2nded.). San Francisco, CA: Morgan Kauffman Publishers.Hand, D. J. (2007). Principles of data mining. Drug Safety, 30(7), 621–622.Lewis, M.(2004). Moneyball: The art of winning an unfair game. New York, NY:W. W. Norton & Company.Lohr, S. (February 11, 2012). The age of big data. New York Times.Sport Analytics58McAfee, A., & Brynjolfsson, E. (October, 2012). Big data: The management revolution. Harvard Business Review.McCue, C. (2005), Data mining and predictive analytics: Battlespace awarenessfor the war on terrorism. Defense Intelligence Journal, 13(1&2), 47–63.Maindonald, J. H., & Braun, J. (2009). Data analysis and graphics using R: An example-based approach. Cambridge, UK: Cambridge University PressManyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C., & Byers, A.H. (2011). Big data: The next frontier for innovation, competition, and productivity. McKinsey Global Institute.Nash Information Services, LLC. (n.d.). The numbers. Retrieved from http://www.the-numbers.com/movie/Moneyball#tab=summaryNew York Times. (2011). Best sellers. Retrieved from http://www.nytimes.com/best-sellers-books/2011-11-20/combined-print-and-e-book-nonfiction/list.htmlPlunkett Research, Ltd. (2014). Sports industry market research. Retrieved fromhttp://www.plunkettresearch.com/sports-recreation-leisure-market-research/industry-and-business-dataR Foundation. (2014). The R project for statistical computing. Retrieved from http://www.r-project.org/Slaton, Z. (February 20, 2013). Why the Sloan Conference is the Super Bowl ofsports analytics. Forbes. Retrieved at: http://www.forbes.com/sites/zachslaton/2013/02/20/why-the-sloan-conference-is-the-super-bowl-of-sports-analytics/Surendra, N. C., & Denton, J. W. (2009). Designing IS curricula for practical relevance: Applying baseball’s Moneyball theory. Journal of Information SystemsEducation, 20(1), 77–86.Taylor, J. (2012). Decision management systems: A practical guide to using businessrules and predictive analytics. Boston: IBM Press.Watson, H. J., & Wixom, B. H. (2007). The current state of business intelligence.Computer, 40(9), 96–99.
Get professional assignment help cheaply
Are you busy and do not have time to handle your assignment? Are you scared that your paper will not make the grade? Do you have responsibilities that may hinder you from turning in your assignment on time? Are you tired and can barely handle your assignment? Are your grades inconsistent?
Whichever your reason may is, it is valid! You can get professional academic help from our service at affordable rates. We have a team of professional academic writers who can handle all your assignments.
Our essay writers are graduates with diplomas, bachelor, masters, Ph.D., and doctorate degrees in various subjects. The minimum requirement to be an essay writer with our essay writing service is to have a college diploma. When assigning your order, we match the paper subject with the area of specialization of the writer.
Why choose our academic writing service?
Plagiarism free papers
Timely delivery
Any deadline
Skilled, Experienced Native English Writers
Subject-relevant academic writer
Adherence to paper instructions
Ability to tackle bulk assignments
Reasonable prices
24/7 Customer Support
Get superb grades consistently
Get Professional Assignment Help Cheaply
Are you busy and do not have time to handle your assignment? Are you scared that your paper will not make the grade? Do you have responsibilities that may hinder you from turning in your assignment on time? Are you tired and can barely handle your assignment? Are your grades inconsistent?
Whichever your reason may is, it is valid! You can get professional academic help from our service at affordable rates. We have a team of professional academic writers who can handle all your assignments.
Our essay writers are graduates with diplomas, bachelor’s, masters, Ph.D., and doctorate degrees in various subjects. The minimum requirement to be an essay writer with our essay writing service is to have a college diploma. When assigning your order, we match the paper subject with the area of specialization of the writer.
Why Choose Our Academic Writing Service?
Plagiarism free papers
Timely delivery
Any deadline
Skilled, Experienced Native English Writers
Subject-relevant academic writer
Adherence to paper instructions
Ability to tackle bulk assignments
Reasonable prices
24/7 Customer Support
Get superb grades consistently
How It Works
1. Place an order
You fill all the paper instructions in the order form. Make sure you include all the helpful materials so that our academic writers can deliver the perfect paper. It will also help to eliminate unnecessary revisions.
2. Pay for the order
Proceed to pay for the paper so that it can be assigned to one of our expert academic writers. The paper subject is matched with the writer’s area of specialization.
3. Track the progress
You communicate with the writer and know about the progress of the paper. The client can ask the writer for drafts of the paper. The client can upload extra material and include additional instructions from the lecturer. Receive a paper.
4. Download the paper
The paper is sent to your email and uploaded to your personal account. You also get a plagiarism report attached to your paper.
PLACE THIS ORDER OR A SIMILAR ORDER WITH Essay fount TODAY AND GET AN AMAZING DISCOUNT
The post business information systems appeared first on Essay fount.
What Students Are Saying About Us
.......... Customer ID: 12*** | Rating: ⭐⭐⭐⭐⭐"Honestly, I was afraid to send my paper to you, but you proved you are a trustworthy service. My essay was done in less than a day, and I received a brilliant piece. I didn’t even believe it was my essay at first 🙂 Great job, thank you!"
.......... Customer ID: 11***| Rating: ⭐⭐⭐⭐⭐
"This company is the best there is. They saved me so many times, I cannot even keep count. Now I recommend it to all my friends, and none of them have complained about it. The writers here are excellent."
"Order a custom Paper on Similar Assignment at essayfount.com! No Plagiarism! Enjoy 20% Discount!"
