Apmis 250) Sujuunya/Sunoys0yy :japoyy uorssasBoy ay, SSA's? 29PL =z zi if 52 = Ret i z a3 |B meme cle ee t cieeeadas Pla ee) =|? 8 8 Eig alg eg 2 3 3 ae Z 4 uence sel gil) 8 4 eee lias 2 8 2 B 2 e 2 a |e Z <3 | 5 2 2 aie e 3 2 a\e| FF 2 z a ale = 8 s = g lee ca sa ee es 5 a ran Biceseeteal ele |) oles) Slal fue Mies a eis & Oo: == ec eis BRIGG ais aa a z aj2 2 Sigs es BE 5 }— = lence s == Ba dies z aie: =| fee Els = -| § reece e cel ply sees 2 al = z SNe lit = 4 g &| 8 z elena g Ss my og = |e & ie alle o¢ Ean BS 2 Sal| 2 \B Sy ese Se . elie Silyeemate BI 8 BE S| Es lE 3 8 eI = Pale g Sas i eee 2 es > 5 2 Fe oa =| § z el le 22 22 Predictive Analytics REGRESSION MODEL USING EXCEL POS Regression Model ‘Summary Output Regression Statistics Multiple R 0.0126486 R Square 0.000159 Adjusted R Square -0,0553867 Standard Error 268855013 Observations 20 ANOVA af ss MS F__| Significance F Regression 1 | 20819.162| 20819.16 | 0.00288] 0.957791029 Residual 18| 130109432] 7228302 Total 19 | 130130251.2 Coetticients| S@4*F4 | i seat | p-vatue | Lowerss% | Upper | Lower Error_| ‘" 95% 95.0% Intercept | 16649.4448 | 1398.322032 | 11.90673 | 5.72B-10 | 13711.67923 | 1958721 | 13711.6792 | 19587.21039 pos _| 4.14019 _ | 822.4712269 | 0.053668 | 0.957791 | -683.807738 | 1772.0881 | -683.80774 | 1772.088118 Table 4.6. Excel POS Regression Model: Marketing/Planning Case Study Soudqouy ssouisngy- Business Analytics Radio Regression Model = Summary Output Regression Statisties Multiple R (0977138075 R Square 0954798817 Adjusted R Square 095228764 Standard Error 571.646807 Observations 20 ANOVA = af ss MS F__| Significance F Regression 1 | 124248209.9 | 124248210 | 380.2197 1A92E-13 Residual 18 | 5882041.294 | 326780.072 Total 19 | 130130251.2 Standard Lower Upper Lower Upper Coefficients eee rstat | P-Value | go, a ae See intercept |= 74192148 | 1362,030419 | -7.1477289 | 1.176-06 | ~2605.35 6878492 | -12605.351 | -6878.49202 radio | a7.csees | 17.83090866 | 19.4992222 | 1496-19] 310.2975 _] 385.150199 310227501 _| 385.150199 Table 4.7. Excel Radio Regression Model: Marketing/Planning Case Study Predictive Analytics TV Regression Model ‘Summary Output Regression Statistics Multiple R 0.9579703 R Square 0.917707 Adjusted R Square 0.9131352 Standard Error 77131951 Observations 20 ANOVA af ss Ms F__| Significance F Regression | 1] 119421443] 1.19E+08 | 200.7306 3,336E-11 Residual 18 1070880822] 594933.8 Total 19 | 130130251.2 Coefficients Standard Error _¢Stat__P-Value Lower 95% Upper 95% Lower 95.0% Upper 95.0% 95, per 95. Intercept 4229.21 4164.121037 10.1412 7.19E-09 0977.7 -33480.714-S0977.702 ~33481 na = 33480.7145 T 221, Vv 1.10431 15.60595543___14.16794 3.34E-11_ 18831741 253.8912 188.317411 _253.891202: . 2023 Table 4.8. Excel TV Regression Model: Marketing/Planning Case StudyBusiness Analytic, 4.48) (a Inference of the results ‘The computer printouts in the tables provide a variety of st for comparative purposes. Diseussion .d here to just a few. The R-Square statistics are g precise proportional measure of the variation that is explained by lependent R-Square t0 1.00, the variable’s behavior with the dependent variable. more of the variation is explained, and the variables’ R-Squares are 0.000 (POS), 0.955 (radio), and 0.918 (TV). Clearly, radip is the best predictor variable of the three, followed by TV and, without almost any ier the predictive variable. The three relationship, POS. This latter result was expected based on the pri jon. What Pearson Sora (000-0.918) of the is suggesting is that only 0.082 percent variation in product sales is explained by TV commercials. actually comparing the regression the dependent variable. As R-Square increases, so docs the Fa because of the way in which they are computed and what is measured by both. The larger the ), the greater the st significance in explaining the variable’s relationships. The three variables’ from the ANOVA tables are 0.003 (POS), 380.220 (radio), and 200.731 (TV). Both but POS has an insignificant relationship. To give some idea of how significant the relationships are, assuming a level of significance where a = 0.01, one would only need a cut-off value for the F- of the case of as being signif ant. Not exceeding that “ratio (as, the same as saying that the coe jent in the regression modi POS is no different from a value of zero (no contribution to Product Sales). Clea the independent variables radio and TV appear to have strong relationships withthe dependent variable, The question is whether the two combined or even Variables might provide a more accurate forecasting model than just using the one best variable like radio. Predictive Analytics S are summarized and presented in Table n solely on the R-Square statistic, there there is a tie Betwesn the FOSMedio/TV and the radia/TW’ coniGination (0979 R-Square values) IF the decision is based solely on the F-ato value fom m, which one might expect of the two most mn and to ensure these analytics are the best possible esti nal statistic can be considered, That tie breaker isthe Re Squared (Adjusted) statistic, which is commonly used in multiple regression models, Variable RS Kea RSquare nore ee Combination (Adjustea) | F-Ratio POS/radio 0.957 0.952) 188.977 Pos/ry 0.920 0911) 97.662 POS/radio/TV 0.979 0951) 123.315 Radio/TV 0979 0.953] 192.555 Table 43. SPSS Variable Combinations and Regression Model Sta Marketing/Planning Case Study Inference ‘The R-Square Adjusted statistic d (@ precise, Proportional measure of comparative measure of suit Joes not have the same interpretations R-Square variation in the relation-shi lity of alternative independent variables, instead a is ideal_— = Business Analytics for selection between independent variables in a m Square adjusted seeks to take into account the phe enon of the R-Square ‘automatically increasing when ad dependent variables are added 10 the 2 paint on a canvas, where more rely increases the value of the pai ‘which some paint covers other paint, diminishing the value of ‘many variables, some of wl predictive ability of the model. The R-Square adjusted. sta information to aid in revealing this behavior. Jess than or equal to that of the R-Square in which itis rel Square, R-Square adjusted increases when a new independent varia is included only ifthe new variable improves the R-Square more than would be expected in the absence of any independent value being added. If a set of independent variables is introduced {mo a regression model one at a time in forward step-wise regression using the high est correlations ordered fi , the R-Square adjusted end up being equal to or less than the R-Square value of the original model. By system: ation with the R-Square adjusted recomputed for each added variable or combination, the value of the R-Square adjusted will reach a maximum and then decrease. The multiple regression mode! with the largest R-Square adjusted stati n of having the best fit without excessive or ‘unnecessary independent variables. Again, just putting all the variables into a mode may add unneeded variability, which can deerease its accuracy. Thinning out th variables is important. 4.5, BEST VARIABLE COMBINATION REGRESSION MODEL Finally, in the step-wise multiple regression procedure, a final decision on the variables to be included in the model is needed. Ba: adjusted, the best combination is radio/TV. The SPSS m support statistics are presented in Table 4.12, and the Excel model is shown in Ta 443. regression model. The R. * Predictive Analytics asi) Model Summary Elly Std. Change Statistics vse] |. peice Errorof |p ware Rsquare| the F jamare | the | Square an}an| S8F Estimate | Change | Chee ‘Change 1] 979%} 958 | 953 | s68:87547| 958 [192555 2 [17 | 000 4, Predictors: (Constant), TV, Radio Coefficients ae Unstandardized | Standardized Coefficients Coctticients |‘ | Sit Std 5 . Ewe | Bete (Constant) =17150.455 | 6.965.591 -2.462| 025, 1 Radio 275.691| 68.728 75|_40u| oor Vv 48341] 44.580 209} 1.088] 293 a, Dependent Variable: Sales ANOVAS Sum of Model Squares | % [Mean Square] Ff | sig. Regression | 124628723.140] 2 | 62314361.570| 192555] 090 1. Residual 5501528060] 17] 323619298 Total 130130251.200 | 19 4a. Dependent Variab. Table 4.10. SPSS Best Variable Combination Regression Model and Staite: ‘Marketing/Planning Case Study’ Business Analytics 20 0.97863319 0.95772291 0.95274914 968.875468 Regression Statistics ‘Adjusted R Square Standard Error Observations Multiple R R Square 4-52) ‘Summary ‘ANOVA predictive Analytics -2454,342976 420.6940765 142,3970308 Lower 95.0% | Upper 95.0% 31846 .5678 130,687223 45.7158836 Upper 95% 2454343 420,694077 142,397031 Lower 95% 31846.56777 130,6872233 -45.11588363, P-Value r Stat 2.46217 | 0.024791 4.011329 | 0.000905 1,084345 | 0.293351 eS e8 €\2 a ola 83 ales Zin 8 gle es sine a3 33 e228 68.72801022 445804165 Coefficients | Standard Error 275.69065 48 3405736 Intercept | -17150.4554 | 6965.590997 Radio 1 Table 4. el Best Variable Combination Regression Model and. -s: Marketing/Planning Case Study M1. Exe ination Regression Model and Statistics: Marketing 18 Ca ay 7 ‘arial [4°53] Inference Although there are many other additional analyses that could be per-formed to validate this model, we will use the SPSS multiple regression model in Table 6.12 for the firm inthis case study. The forecasting model can be expressed as follows: Yp = -17150.455+275.691 x, + 48.361 x, where: > = the estimated number of dollars of product sales the number of dollars to invest in radio commercials the number of dollars to invest in TV commercials Because all the data used in thie model is expressed as dol , the interpretation of the model is made easier than using more complex data. The interpretation of the multiple regression model suggests that for every dollar allocated to radio commercials (represented by X,), the firm will receive $275.69 in product sales (represented by Y pin the model).Likewise, for every dollar allocated to TV commercials (represented byX,), the firm will receive $48.34 in product sales A caution should be mentioned on the results of this case study, Many factors might challenge a result, paicularly those derived from using powerful and complex methodologies like multiple regression. As such, the results may not occur as estimated, because the model is not reflec-ing past performance. What is being suggested here is that more analysis can always be performed in questionable situations. Also, additional analysis to confirm a result should be undertaken to strengthen the trust that others must have in the results to achieve the predicted higher levels of business performance. Conclusion In summary, for this case study, the predictive analytics analysis has revealed a more detailed, quant ble relationship between the generation of product sales and the sources of promotion that best predict sales. The best way to allocate the ©350,000 budget to maximize product sales might involve placing the entire budget into radio commercials because they give the best retum per rupee of budget. Unfortunately, there are constraints and limitations regarding what can be allocated- 4 Business Analy, Predictive Analytics to the different types of promotional methods. Optimizing = aon of resource and maximizing business performance pena peat business analytic methods designed to accomplish this tas requires the additional step of prescriptive analytics analysis in the BA process. ‘SHORT QUESTION WITH ANSWERS 1. Define predictive analytics. We defined predictive analytics as an application of advanced statistical information software, or operations research methods to ident and build predict observed in the deseript explains why one set of independent variables (predictive variabl. ce business performance. 2. List out the predictive analytics techniques. Data Mining es models to identify trends and rel ic analysis. Knowing influences dependent variables ical Modelling, ‘ “Machine learning 3. What is predictive modelling? Predictive modeling means the developing models that can be used to forecast ‘or predict future events, Models can be developed either through logic or data, Write a short note on logic-driven models. Logic driven models remain based onl experience, knowledge and logical relationships of variables and constants connected to the desired busines: performance outcome sit ion. 5. What is data-driven models? Data-driven Models refers to the models in which data is collected from many ‘Sources to qualitatively establish model relationships. Logic driven model: ‘often used as a first step to establish relationships through data-driven models Data driven models include sampling and estimation, regression analysis ‘correlation analysis, forecasting models and stimul 6 10. 455) What is predictive analysis modelling? Predictive modeling and analysis might consi including those inference, data Of the use of methodologies, 16 and estimation, statistical i, and regression analysis. A commonly used methodology found in forecasting, samp . Name some common predictive algorithms, Random Forest: Generalized Linear Model (GLM) for Two Values: Gradient Boosted Model K-Means: © Prophet Write a short note on Data mining. Data mining is discovery-driven software application proces that provides his into business data by finding hidden patterns and relations small data and inferring rules from them to predict future behavior observed pattems and rules guide decision-making. Th text and social med in big or information from the Web. Data mining is an ‘ideal redictive analytics tool used in the BA process What are the data mining algorithms? K-means % Apriori Algorithm % K Nearest Neighbor % Naves Bayes # AdaBoost ‘Name some data extraction methods. % CLIQUE * PIC & Lac * CURLER— en Business Ano| 11. List out the uses of predictive analytics in business. + * + + * Improve Customer Retention Identify Profitable Customers Improve Customer Segmentation Improve Decision Making Perform Predictive Maintenance 12. Mention the benefits of predictive analytics. * ° ° ° 13. Name the appli ° eo ee oe ee * Detecting Fraud Reducing Risk Optimising Marketing Campaigns Improving Decision Making Improving Efficiency in Operations ion of predictive analytics in various industries. Banking and Financial Services Debt Collection Insurance ‘what are the method used in logic driven model? Diagramming methods Influence diagram 14, List out the types of data-driven models, Predictive Analytics eee ee (47) ——— Ee 15. 16. 17. 18, 19, Predictive Modeling and Analysis ‘Forecasting Models > Simul n What are the Types of Predictive Models? > Class mn model Clustering mode! Forecast mode! Time series model Write short note on Predictive modeling Predictive modeling is also known as predictive analytics. General, the term “predictive modeling” is favored in academic setings, while “predic analytics” is the preferred term for commercial applications of predictive modeling, Narrate the Benefits of Predictive Modeling, In a nutshell, predictive analytics reduce time, effort and costs in forecasting business outcomes. Vi intelligence, reg mathemat es such as environmental factors, competitive n changes and market conditions can be factored into the ion to render more complete views at rl calcul What are the limitations of predictive modelling? 4 Errors in data labeling Shortage of massive data sets needed to train machine learning, ‘The machine's inability to ex % Generalizability of learning, or in data and algorithms List out the predictive modelling procedure. ‘Defining the problem statement/business goal. % Col- (ess) Business Analy, n. Integrate the data obtained from various sources. Analysis of data with analytics tools/software. - Model Generation + ce aay 4 Validate assumptions, hypotheses and test them using statistical models, + in 4 Deploying the model to generate predictions and monitor them fo, accuracy. |. What is Classification Model? mne of the simplest of all the models. It categorizes new data based on fication by answering binary questions like Yes/No, True/False but can be used for multiclass classification as well. Decision Trees, Support Vector Machines are some classification algorithms. |. What is Clustering Model? A clustering model sorts data points into groups based on. simil attributes. There are many clustering algorithms but no algorithm can be claimed as the best for all use cases. It is an unsupervised leaning algorithm, unlike classification which is supervised. What is Forecast Model? Being one of the mi idcly used predictive analytics models, it metric value prediction, estimating a numerical value for new data based on the Jeamings from historical data. It can be applied wherever numeric data is available. Eg. Traffic prediction at a city's main road during different periods. Stores estimating availability of products in their warehouse. Write a note on Outliers Model. As the name suggests, it is based on the anomalous data entries in their dataset. An outlier could be a data entry error, measurement error, experimental error, intentional, data processing error, sampling error, or natural error. Even though outliers can cause poor performance and accuracy, some help us find the novelty or to observe new inferences, oe 24. What is Time Series Model? future data using that metric 25. List ou the Merits of Data Mining. + Data mining technique hel information, Ps companies to get knowledge-based Data mining helps organizations 10 make the profitable adjustments in operation and production. 8 'S a costeffective and efficient solution compared 10 data applications. helps with the decision-making process. Facilitates automated prediction of trends and behaviors as well as ‘automated discovery of hidden patterns. {can be implemented in new systems as well as existing platforms tis the speedy process which makes it easy for the users to analyze huge ‘amount of data in less time. REVIEW QUESTIONS, — $$$ — Describe the uses and importance of predictive analytics in business. What are the benefits of predictive analytics. Outline the uses or application of predictive analytics in various industries. Explain the process involved in predictive analytics. Discuss the need of predictive analytics to attained key strategic objectives of firm. Explain the different method used in logic driven model 7. Narrate the types of data-driven models. Elaborate the types of predictive models.: Business Analy, 4.60 . 9, What are the limitations of predictive modelling? 10. Explain the predictive modeling procedure. . Describe about the predictive analytics models. . Explain the data mining process. 13. Outline the merits and demerits of data mining. “uss about various data mining methodologi ications in various field. 15. Comment on data mining apy 3s step in the BA process with suitable example, 17. Explain what logic-driven models are used for in business analytics (BA) 18. Describe what a cause-and-effect diagram is used for in BA. 19. Explain the difference between logic-driven and data-driven models. Explain why neural networks can be helpful in determining both associations and classification tasks required in some BA analyses. 22. Explain how clustering is undertaken in BA. Explain how step-wise regression can be useful in BA. Explain how to use R-Squared adjusted statistics in BA. . Why is predictive analytics analysis the next logical step in any business analytics (BA) process? Why would one use logic-driven models to aid in developing data-dri models? . How are neural networks helpful in determining both associations and classification tasks required in some BA analyses? Why is establishing clusters important in BA? 29. Why is esta ing associations important in BA? 30. How can F-tests from the ANOVA be useful in BA? UNITV PRESCRITIVE ANALYTICS © Modeling - Non Linear Business Performance Improvement. leps in the BA process, one should be positioned to undertake the final step: prescriptive analytics analysis The prior analysis should provide a forecast or prediction of what future trends in the business may hold. For example, there may be significant st increased (or de-creased) sales, profital new market opportuni tends accurately measured in dollars for s, ot measured cost savings from a future venture. Ifa firm knows where the future lies by forecasting trends, an best plan to take advantage of possible opportunities that the trends may offer. Step 3 of the BA process, prescriptive analytics, involves the application of decision science, management science, or operations research methodologies to make best use of allocable resources. These are mathematically based methodologies and algorithms designed to take variables and other parameters into a quantitative framework and Generate an optimal or near-optimal solution to complex problems. These methodologies can be used to ed resources to take best advantage of the opportu has found in the predicted future trends. Limits ‘on human, technology, and financial resources prevent any firm from going after all the opportunities. Using prescriptive analytics allows the firm esources to optimally or near-optimally achieve the objectives as fully as possible. 5.1.1. MEANING timally allocate a firm's Prescriptive Analytics bridges the gap between organization and the associated im data possessed by an it for various purposes. It ms of explo- 5 Business Analytics enables end-users of existing information systems to connect within ay organizational ecosystem. Prescriptive Analytics can also help decision-makers in any organization get an understanding of how they can take advantage of a future ‘opportunity or alleviate future pain points. Overall, it reduces the risk of making decisions. Prescriptive analytics is a process recommendations on how to optimize business practi ‘outcomes. In essence, prescriptive analytics takes the ‘comprehensively understands that data to predict what could happen, and suggests the best steps forward based on informed simulations. data to determine an optim: analyzes data and” provides instant Prescriptive analytics is the process of usi of action. By considering all relevant recommendations for next steps. Because of this, prescriptive analytics is a valuable tool for data-driven decision-making 5.1.2. BENEFITS OF PRESCRIPTIVE ANALYTICS Today, businesses are using data as a driving force. Therefore, the importance of Prescriptive Analytics is well-known. It will not only give you valuable insights from your data but will also optimize your data-driven decisions. The, key benefits of Prescriptive Analytics are as follows: ‘More Proactive: ‘The fundamental idea behind transitioning from a predictive model to to empower organizations to become more proactive and less Prescriptive Analytics can address issues found during the Pre: ‘Analysis and make weighted decisions for better business outcomes. Capturing Multiple Data Touchpoints and Formats: Prescriptive Data Analysis models are built on input data, business rules, and mathematical models, sourced within the organization (internal data) or collected from extemal touch points. The gathered data can cither be Structured ot Unstructured (like sensory or audio-visual data). Businesses can use Prescripi ‘Analytics to connect the dots between different data sets to garner insights i © events, This allows Prescriptive Analytics to be used Prescriptive Analytics in industries like healthcare and automobiles for business growth, Real-Time Insights: Enterprise leaders can Possible that there are multiple solutions available to a given problem. ‘tricky and overwhelming process, is by running complex models, anal ls, analyzing, ing scenarios by factoring in business rul iness rules and constrai re best among those solutions, al Maximum Use of Resources: Prescriptive Analyti ; helps organizations make costeffective decisions by effectively managing resources, thereby minimizing the need to outsource. This is achieved by enabling employees with minimal knowledge to learn about data and the impact of data-driven decisions. It also allows them to participate or co ‘on data-specific processes. Gross Margin Management: Since Prescriptive Analytics is more actionable than its predictive counterpar, empowers developers to build solutions based on the current ‘anticipating market conditions and customer behavior teams can spend more time designing the perfect so problems. Enhanced Market Competition Analysis: Brands can run diffe Anal ed experience than their competitors. Using these analysis iscover and. offer a superior person:Prescriptive Analytics powered personalized recommendations, marketers can view plans that help them, significantly impact the market while resonating with customer demands. Removing Bottlenecks: By identifying pain points, Prescriptive Analyties can also help solve issues in he 5 model, they wi actions in neartme a opposed wecks or me ne i hei cks or months | or iy changed. later when market condi Create a repeatable, scalable process: meet pia To accurately model complex scenarios for prescri Experience cost efficiencies with in-house capa iW’s necessary to create an accurate twin of the marke With self-service analytics tools such asa simulation prescriptive analysis, businesses save money and also mak {As opposed to outsourcing their analyt : ve decisions. ive analytics, prescri and predi By building off descriptive, diagnos i ions take into consideration historical appl ta and forecast 1 to resulls businesses have a repeatable jon making ternal departments who analytics to be involved in a collaborative process and act of their decisions through a user-friendly platform, to what occurred in the market to validate the model process that they trust to help with their d Optimize business actions: ‘The reason Forbes predicts that the future of data analytics is pres analytics is because of its al business. The process brings teams together to collaborate who may not converse ing What will happen in an f business matters otherwise but strategic decisions n pricing, business users are al on also allows departments to focus their efforts on their expertise, Self-service analytics are designed to be user-friendly so businesses Users are empowered to run scenarios to help them determine what to do next. This ‘means IT departments and data scientists continue to dive into the data as business users dor reduces productivity. organization, but how it could happen beter by making certa Whether egy or chi n {0 choose the best course of understanding what could ‘with recommendations for how ve their goals. By not on happen in the future but actual get there, businesses are able to optimize their actions in stable economic conditions los and communication blocks so every department improves their 1e5 of uncertainty. E 5.1.3. REAL-WORLD APPLICATIONS OF PRESCRIPTIVE ANALYTICS IN BUSINESS Use near-time decision-making: Perspective Analytics is a process that can provide valuable contributions to Various business ecosystems. Due to this property, it can be applied to a number of this agile and accurate model in place, business users at all levels of the minutes. Since the mode! has bee jon are able to run scenarios in mi- &) Business Analytic =—S Ss a Prescriptive Analytics different business markets. The real-world applications of Prescriptive Analytics ay. Ws oscnrca as follows: ‘> Banking, Financial Services and Insurance (BFSI) © Same alg a pitches as per customer preferences, 'onthms to determine pricing and sales Healthcare + 5. Supply Chain and to Online Leaming SiS crucial for route optimi companies leverage shipping locations. They also rely on min the Supply Chain 10 prevent logistical issues ‘> Transportation and Travel ia i sues energy consumption while saving time and mo 6, Manufacturing © incorrect ails lyse for better route planning Supply Chain and Logistics mney. Manufacturing lesser Marketing and Sales Factories can ana Ze real-time data with Prescri 1. Banking, Financial Services and Insurance (BFS!) iptive Analytics for enhanced tion management This includes predic Financial institutions can design Prescr es algorithms for managing risk and profitability by sifting through historical trading data. Some insurance ‘companies also employ risk assessment models to provide better premium information about insurance policies for clients. 2. Healthcare l. Prescriptive complying with the Prescriptive Analytics s doctors to make data-backed decisions ang treatment recommendations based on the medical history of patients. Apart from assessing risk magnitude, these analyses also enable them to determine the bes action plan and even measure the efficacy of discoveries. It even helps identify optimum setings to increate yield while being efficient. 7. Marketing and Sales Since Presc ferventions. And for hospital admins this analysis can assist in improving clinical care, scheduling treatments, and follow. up appointments. © Analytics is more about modeling than experimenta Perfect asset for brands looking to strengthen their Marketing techniques, helps run promotional campaigns and forecast demands with respect to segments’ ‘consumption and customer interests, 3. Online Learning Numerous Leaming Management Systems (LMS) and websites leverage 5.1.4. STEPS INVOLVED IN PRESCRIPTIVE ANALYTICS Pres Determine the problem to be solved: findings from those tests, LMS presents them with a perso: 4. Transportation and Travel Establish a first use case, preferably one that's not overly complex and where there's a reasonable prospect of an early success. Obtain executive support Im airline companies, Presc ‘Verify the organization is prepared to sponsor and support a proof of concept case, especially in terms of engaging with a vendor, redeploying staff and establishi a budget(53) Business Analy, Select the right technology: Determine how the model is going to be prepared, who's going to do the work any what technology to use. Engage a vendor, or if you're going to hardcode the moc from scratch, appoint a data scien Determine the scope of the project: Est sh who would be involved, whether a limited model would suffice ory there's a need to obtain support and commitment from multiple business users for large-scale project. Establish a vision: Determine what you hope to achieve. Set up a team: Establish a team including a sponsor, project leader and ste ‘committee, Project design: Determine model requirements and refine the scope of the project Build and validate the model: Create a model that represents the problem, populate it with known data, the model to ensure it accurately represents known outcomes. Prepare your data: and val Identify all data required and transform data as necessary so the model can read anj use the data. User interface: Set up an interface that allows users to enter data, run scenarios and inter, results. Test the modi Comprehensively test the model to ensure it works as required and that ironed out, G Implementation and ro tart using the model. 5.1.5. ADVANTAGES OF PRESCRIPTIVE ANALYTICS 1. Generation of revenue Prescriptive analytics applications can provide de information about the customers’ preferences, identify new opportu also allows business managers to selling and accelerating the regular sales lies for cross-s Prescriptive Analytics levels of revenve. 2. Management of Grosé Margins expenses better 5.1.6. LIMITATION / CHALLENGES OF PRESCRIPTIVE ANALYTICS * The impact of bad decisions escalates faster with prescriptive analytics, An ting from automated trading. Because of the stringent data engineering requirements of prescriptive anal ions may not be feasible for using this type of 4 checkout app using prescr make customers wait too long Bias can be ac ly or intentionally Perpetuated by automation, ptive analytics might ly baked into analytics models and is to explain how the model works, can be 'S models, putting companies at risk of ns such as GDPR. © Performance infrastructure and processes for Ongoing monitoring and adjustment 8 social media algorithm might promote 's shared more widely. In some cases, viral content driven . rs would need to consider ways to mitigate this ded consequenceBusiness Analytic, prescriptive Analytics 5:10) 15.1.7. THE DISTINCTION BETWEEN PRESCRIPTIVE ANALYTICS AND PREDICTIVE ANALYTICS , ric We have seen how prescriptive and predictive a help businesses generate profit and reduce expenses. However, there are Prominen differences between the two concepts. s work in combination tg Predictive Analytics Imagine you go to a fortune teller and learn what the future bolds for you in many ways. However, the objective of predictive wre, No one can do that. However, predictive analyi if a cyclonic storm is brewi analytics forecasts the probability ofthe eyclone striking land at a particular time, Prescriptive Analytics Taking the same example forward, prescriptive analytics applications help to fing solutions to withstand the cyclone and ensure that it causes minimal harm to both and property. It prepares to face the hurricane. It cannot combat it or prevent it from happening, but you can take steps to minimize damage. Prescriptive analytics focuses on these aspects of mitigating and facing risk. 5.1.8. PRESCRIPTIVE ANALYTIC METHODOLOGIES Integer programming Nonlinear optimization Other methodologies Simulation Fig. 5.1. Prescriptive analytic methodologies St The listing of prescriptive analytic methods and models in Figure 5.1 is but a small grovPing of many operations research, decision science, and management sefence methodologies that are applied in this step of the BA proces 3S. 5.2. PRESCRIPTIVE MODELING = ee The name ‘prescriptive’ is given because the model prescribes «feet et eed vines actions, tasks, quality assurance and change the mechanism for every Project. ' There are three types of prescriptive process models. They are ‘The Waterfall Model 4 Incremental Process model 4% RAD model 5.2.1. THE WATERFALL MODEL ‘The waterfall model is also called as ‘Linear i ! oF ‘Classi aes sequential mode!’ or ‘Classic % In this model, each phase is fully i completed before the beginning of the ‘This model is used for the small * ojects In this model, feedback is taken after each is on the right path, phase to ensure that the project ‘® Testing part starts only after the development is ‘complete. + Estimation Fig. 5.2. The Waterfall ModetBusiness Analy “|___ aes ‘The description of the phases of the waterfall model is same as tha op the process model. k Vis 3: ‘An altemative design for linear sequential mode!'is as follow Fig. 5.3. The linear sequential mode! Advantages of waterfall model 4 The waterfall model is simple and easy to understand, implement, and use 4 All the requirements are known at the beginning of the project, henee it easy to manage. Itavoids overlapping of phases because each phase is completed at once 4 This model works for small projects because the requirements are understood very wel. ‘ This model is preferred for those projects where the quality is more important as compared to the cost of the project. Disadvantages of the waterfall model + This model is not good for complex and object oriented projects. Itis.a poor mode! for long projects. + The problems with this model are uncovered, until the software testing, The amount of risk is high 5.2.2" INCREMENTAL PROCESS MODEL + The incremental model combines the elements of waterfall model and they are applied in an iterative fashion. % The first increment in this model is generally a core product. % Each increment builds the product and submits it to the customer for any suggested modifications. prescriptive Analytics a + The next increment implements on the customers suggestions and add additional requirements in the previous increment. “This process is repeated until the product finished. For example, the word-processing software is developed using the incremental model. QO communication © Paring @ Meeting (anahysis, desing) B consiucion (code, est) @ ds sicyment seivary feedback) ncrement 3 sn OO coteyet screen ecromens DBD over 2 ircement DO ener at nincenen Project calender time Sofware functionality and features Fig. 54, Incremental Process Model ‘Advantages of incremental model ‘> This model is flexible because the cost of development is low and initial prodi ‘> Iti easier to test and debug during the smaller iteration, +> The working software generates quickly and early during the software life cycle. 4 The customers can respond to its fune ties after every increment. Disadvantages of the incremental mode! + The cost of the final product may cross the cost estimate intially. 4 This model requires a very clear and complete planning, The Wg of design is required before the whole system is broken into small incrementsa) Business Analytics Prescriptive Analytics 4 The demands of customer for the additional functi increment causes problem during the system architecture. ies after every 5.2.3. RAD MODEL + RAD isa Rapid Application Development model. 4 Using the RAD model, software product is developed in a short period of time. + The initial activity starts with the communication between customer ang developer. Planning depends upon the initial requirements and then the requirements, are divided into groups. % Planning is more important to work together on different modules. The RAD model consists of following phases: Phase 1 Business Modeling 4 Business modeling consists of the flow of information between various functions in the project. produced by every function and information. A complete business analysis should be performed to get the essential business information. Phase 2. Data modeling % The attributes of each object are identified and define the relationship between objects Phase 3. Process mod * The data objects defined in the data modeling phase are changed to fulfil the information flow to implement the business model. (5] ‘® The: process description is created for adding, modifying, deleting or retrieving a data object. Phase 4. Application generation In the application generation phase, the actual system is built. % To construct the software the automated tools are used, Phase 5. Testing and turnover ‘The prototypes are independently tested after each iteration so that the overall testing time is reduced, ‘® ‘The data flow and the interfaces between all the components are filly tested. Hence, most of the programming components are already tested. Fig. 55. RAD Model 5.3. TYPES / SELECT PRESCRIPTIVE ANALYTIC MODELS Linear Programming A. general-purpose modeling methodology is applied to strained, complex multiple ering media, allocatingBusiness Analyticy wuman mology resources 10 product production and optimizing blends of human and tec ‘mixing ingredients to minimize costs of food products aa the same as LP, but it permits decision variables to be integer values Eee include allocating stocks to portfolios, allocating personnel to jobs, an allocating types of crops to farm lands Nonlinear Optimization ‘optimal or near optimal solutions Examples include solving for o ‘making ‘environment for a firm whose analytics have predicted growth in the near future. ‘Simulation This methodology can be used in prescriptive analysis in situations where Parameters are probabilistic, nonlinear or just too complex to use with other ts prescriptive Analytics ay Others Methodologies The areas of operations research, decision sciences, and ie management science combine the foricaion of mathematics, eagicerng, an computer cites offer 5.4. NON LINEAR OPTIMISATION or 5.4.1. MEANING Linear Programming (LP) is a deterministic, muhivarible, constrained, single objective, optimization methodology. It’s a model with known, deterministic, and ‘Constant Parameters, and it has more than one unknown or decision variable. LP has mathematical expressions that constrain the values of the decision variables, and it Seeks 10 solve for an optimal solution wih a single objective. It is a general-purpose Linear programming or linear optimization is a process that takes into ‘consideration certain linear relationships to obtain the best possible solution to a mathematical model. It includes problems dealing with maximizing profits, ‘minimizing costs, minimal usage of resources, ete. These problems are known “%slinear programming problems (LPP). The LPP’s applications can be found inBusiness Analytics tribution, military, broad disciplines such as commerce, industry, marketing, dis i LP is considered for use in the ‘economic, business, ete Modeling a problem using LP is called programming. As f several mathematical programming methodologies = fone of seve prescriptive step of the business analytic process 5.4.2. TERMINOLOGIES / CHARACTERISTICS OF LINEAR PROGRAMMING PROBLEMS Objective Function hat sh Inis defined as some numerical valve t ‘example, if you are involved in some business, then your primary go: ‘maximize profits and reduce loss. be maximized or minimized. For Constraints ‘The constraints are defined as the jions of the decision variables. For , then the budget, number of workers, ‘example, if you are involved in some bu product Decision Variables Decision variables are the variables that decide the output. For example, if you are a farmer who wants to grow wheat and barley, then calculating the total area for growing wheat and barley are the decision variables. ‘capacity, space, etc. are the limitations or restrictions. Non-negativity Restriction ‘Non-negativity jon means that the values for decision variables should be ‘greater than or equal to 0. Finiteness ‘There always should be finite and infinite input and output’ numbers. If the function has infinite factors, the optimal solution will not be feasible. XTIONS OF LINEAR PROGRAMMING PROBLEMS ean be used to model problems in all the’ functional areas of ig, Economics, Finance, Management, and Marketing) and in all ‘pes of operations (industry-wide, government, agriculture, health care, and so on). prescriptive Analytics : (a9) <_< eee Linear programming is used to find optimal solutions for operations research. Linear programming requires the creation of inequalities and then graphing those to solve problems, Here’s a list of areas where linear programming is used. Marketing Management Linear programming allows marketers to analyze the audience coverage of advertising based on constraints such as available media, advertising budge, etc. Linear programming also helps the salespersons (field agents) shortest route for their destination. A logistic head can easily find the optimal distribution schedule for transporting the product from different warehouses to various market locations in such a manner that the total transport cost is the termine the minimum, Financial Management Linear programming helps the financial firms, mutual fund firms, and banks to select the investment portfolio of shares, bonds, etc. in such a manner that the return on investment is maximized, Inventory Management Linear programming also helps the inventory firms in the better management of raw materials and finished products. They can find the optimal solution to minimize the inventory cost based on space and demand as constraints. Human Resource Management Linear programming allows the recruiting manager to solve the problems related to recruitment, selection, training, and deployment of the workforce to different departments of the firm. Linear programming can be used to determine the minimum number of employees required in various shifts to meet the production schedule within a schedule. Food and Agriculture Management Linear programming helps farmers to determine which crops to grow, in what uantity the particular crops should be produced to increase their revenue.- Business Analytics (520! prescriptive Analytics Other Areas Economics: Consumer theory / supplier theory Finance: Optimal hedging / pricing Government: Other Personal decisions: Sports, on-field decisions, player acquisition marketing 5.4.4, ADVANTAGES AND USES OF LINEAR PROGRAMMING “The advantages of linear programming are as follows. 10 business problems. + Linear programming provides insights dimensional problems. 4 Ithelps to solve mul 4 According to change of the conditions, linear programming helps us in adjustments, © By calculating the profit and cost of different things, Littear programming also helps to take the best solution. 5.45. DIFFERENT TYPES OF LINEAR PROGRAMMING PROBLEMS ‘The different types of linear programming problems are: ‘Manufacturing problems % Diet Problems 4 Transportation Problems Optimal Assignment Problems ‘Now, fet us discuss all these problems in detail with their constraints and obj function. ‘Manufacturing Problems Manufacturing problems concern maximizing the production rate or net pr ‘manufactured products, which may be a function of the available workspace, umber of WOTKES, machine hours, packing. material required, the product's market worth, and so on, The = , se sector and can be used to forecast a company’s potential cena Constraints ~ Factors such as labour hours Used in the industrial tal rise over time. : the cost of packing materials, and so ‘The production rate isthe objective function, jet Problems AAs the name implies, diet problems include maximizing the intake of speci Jan. A diet problem's purpose is to find a set of foods that er utrtional requirements for the least amount of money. I meet a set of daily Constraints ~ The nutritional needs it must be met, such as a certain calorie iolesterol in the det. The cost of food consumption isthe objective function ‘Transportation Problems intake or a certain level of sugar or The transportation problems are connected routes, or how e! the study of effective transportation products from various sources of production are carried 10 various markets in such @ way that the ° large organizations with mi analyzing such issues is transportation cost is minimized, For le production units and a large customer base, ical Constraints ~ The unique pattems of supply and demand. ‘The transportation cost is the objective function. ‘Optimal Assignment Problems. The optimal assignment problems are specific task/assignment by se ‘assignment within the required deadline, granted that each individual works on only lated to. a company’s complet g a specific number of employees to ‘ons job within the assignment. Such issues are seen in event planning and management in large corporations, among other places. Constraints — The number of employees the number of hours each employee works, and so on,[S22 Business Analy a The to number of tasks completed is the objective function. 5.4.6, TYPES OF LINEAR PROGRAMMING PROBLEMS SUMMARY Type of Linear Constraints Programming Problem Funetion Manufacturing problems | Variables such as the cost of | Production rate —| Diet Problems Cost of food consumption Transportation problems ‘Transportation cost Optimal Assignment | Work hour of each employee, | Total number of problems imber of employees, and so on | tasks completed 5.4.7. LINEAR PROGRAMMING PROBLEM / MODEL FORMULATIONP. ROCEDURE / STEPS Determine the type of problem A problem has to be is making profit or me problem only mentions cost problem includes sales and and derive profit. Maxi values that can be used to determine formation? Then subtract the cost fron both maximizes sales and minimizes cost, The c type of problem are called the contrt coefficients. Define the decision variables Step 1 det coefficients. The number of pro .ed the type of problem by finding profit or cost contribution or cost contribu ive Analytics There are to things to remember in df what the decision variables are determinin problem is requiring, ning decision variables: (1) Make clear (2) State any “time horizon” the formulate the objective function Because the contribution coefficients, thet decision variables in Step 2 have been identifi into the form of an obj 'yPe of problem in Step 1, and the ied, a left is to combine these ive function formulate the constraints This step is one of the hardest. Here are two strategies that ean hel side strategy: Look available resources the probe =e Problem for a sentence or a column in a table that lists the the model needs to achieve. These are the right-hand-side © & column vector (a column of numbers) that will represent the vo” values in the model. Then go back and read the problem again to find the technology coefficients to finish the left-hand-side of the constraint. (2) Leftchand- side-strategy: In problems with tabled values, look to see if they are technology coefficients. Take the technology coefficients and align them by row or column to form the left-hand.side of the constraints. Then go back and read the problem again to find the right-hand-side values. State the non negativity and given requirements The decision variables in LP models are required to be zero or some positive value. As a formal part of the correct way of formulating an LP model (as is the case in most of the mathematical programming methods),one must add an additional statement in LP model formulations 5.5. NON LINEAR OPTIMISATION 5.5.1. MEANING NONLINEAR PROGRAMMING In many optimization models the objective and/or the constraints are non! functions of the decision variables. Such an optimization model is called a non! programming (NLP) model.or Business An ‘A nonlinear programming (NLP) objective or at least one of the const variables, An example ofa smooth nonlinear function is: 2X? +X} + logX, linear optimization problem is one in which is a smooth nonlinear function of the d ‘When business performance cost sim i models to be useful, exph practice in BA. Although the predictive nature of expression to denote a trend or establish a forecast f analytics step of BA, the use of the nonlinear function to optimize a decision can in the prescriptive analytics step. i functions become too comple ar functions is a standoy loring for a mathemai ‘There’are many mathematical programming nonlinear methodologies and s further complicated in BA by large data files that should be factored into the model-building effort. To overcome these i regression software can be a jons and be more inclusive in the use of large dit, 5.5.2. ADVANTAGES OF NON LINEAR OPTIMIZATION, % Although identical optimization algorithms are used, the NLP proce ‘can be much faster because of the interactive and more general n the IML product, ‘% Analytic first- and second-order derivatives can be computed with : Additional optimization methods are available in the NLP procedure do not fit into the framework of this package. Data’ set processing is much easier than in the IML procedure, You cx save results in output data sets and use them in f + The printed output contains more information. 5.5.3, EXAMPLE OF NON LINEAR OPTIMIZATION, SPSS Curve Fitting software can be used to generate pre ae i Prescriptive Analytics 5.25] Suppose that a resource allocation decision is bei decide how many computer servers a servic f firm's costs of running the facility, The firm's pre analytics effort has shown a growth trend, A new facility is called for it costs can be min The firm hi history of setting up large and small service a ing faced whereby one must should purchase to it best minimizes the distance from the data items io the line, The software then converts the Hine ito a mathematical expression useful for forecasting Data and SPSS Curve Fitting function selection window The first step in using the curve 'g methodology is to generate the best-fiting ‘SPSS models Fig. 5.6. Data and SPSS Curve Fitting function selection window5.26 | U-shaped function/ Cost function In this server problem, the basic data has a u-shaped function, as presented iy Figure 5.7. This is a classic shape for most cost functions in business. In q represents the balancing of having too few servers (resulting in a probles loss of customer business through dissatisfaction and complaints with the service) o, too many servers (excessive wast servers). Although this is an over! investment costs as a rest of underut ed le and nicely ordereg data for clarity purposes, in big data situations, cost functions are considerably les, ‘obvious. 2000.90 a ° 200000) © ° & ° os 7 20000.00 7 8 ° 1500.00 2 . ° 1000080 5 ° eee 5000.00: 00° 0 sb Woo, 1800 «2000 No.of Sree Fig. 5.7. Server problem basic data cost function Adjusted R-Square Values of All SPSS Models The software applies, each point of data using the regression process ‘minimizing distance from a line. The result is a series of regression models an ing ANOVA and other testing statistics. It is known from the regression that the adjusted R-Square statistic can reveal the best ¢ relationship between the in-dependent (number of servers) and dependent (total cos variables, These statistics are presented in Table 5.1, The best adjusted R-Square value (the largest) occurs with the quadratic model, followed by the cubic mod The more detailed supporting statistics for both of these models are presented Table 5.2, The graph for all the SPSS curve-fitting models appears in Figure 5.3 ated re Analytics S27 Lee uray x [RSwer |Add] Stara) (Pome Sr Ee the Estimate IR Square (eae R] Std.Enor of 054 | er250-| [er | ei He | te Estimate The independent variable is Servers =| Logarithmic Medel Summary 7 sane R |RSauare4ste9R] Sis SS fee eee See] Stier Sear | pecan salon stmate | [a01[ 15 | 09 ‘ oz | 8376.020 a Inverse Model Summary hr R [RSquare [4d R] sid Evor ot ac cease Sa ee Square | the Estate al oor | = a Ey | _s7 Quadratic Model Sunmary aa! R [RSavare [dusted] suena lms eee | Gawel |e te Suuve_| te Estimate s| on | - ws | ot | 9 | asans a oss (Cubic Model Summary poten, A | Sware|MsodR] Sizemrar] |" |*S*"| ‘Sees | Seat Square | neesinae| [os] on a ges [932 | 919 | 2sox009 at Compund Model Surmary R [RSquare|AdustedR] Std rer ot Square | the Estima a2s| 001 | 055°] arr Table 5.1, Adjusted RSquare Values of Al SPSS Models Quadratic and Cubic Model na R Adjusted R | Std, Error of the Square | square Estimate [965 | 931 m3 242.315 The independent variable is ServersBusiness Analytic, SS Sum of Squares| df | Mean Square | F__| Sig. Regression | 1266704838.323 | 2| 633352419.161 | 115:440 | .000 Residual 93269442.877| 17| 486437.816 Total 1359974281.200| 19 Coefficients, Unstandardized Standardized 1 Sig. Coefficients Coefficients B Std.Error| Beta Servers =5589.432| 382.188 -3.909| 14.625) 000 Servers**2 268.449| 17.678 4.058) 15.186] .000 (Constant) 3541772 | _1742.639 20.324) —.000| Cubic Model Summary R R__ | Adjusted R | Std. Error of the Square | square Estimate 965 | 932 219 2404,009 The independent variable is Servers ANOVA ‘Sum of Squares| df | MeanSquare | F | Sig. Regression | 1267506094.832| 3 | 422502031.611 | 73.107 | .000 Residual 92468186.368| 16 | 5779261.648 Total 1359974281.200| 19. Prescriptive Analytics re Coefficients dardized | Standardized 1 i Coefficients Coefficients oe B__|siaerror| bem _ | Servers | -5954.738] 1056596 4.164] 5.636] 000 431 4.700] 2.693] 016 Servers**3 1.347 3.619 Servers**2| 310.895 =399} -3r] ms] (Constant) | 36133.696| 2625.976 13.760] 000 Table 52. Quadratic and Cubic Model SPSS Statistics Graph of all SPSS curve-fitting models ee 20000,00} 10000,00- S00 1000 158.00 +2000 No, of Servers Fig. 5.8. Graph of all SPSS curve-lting models ing two statistically significant curve-fited models Y = 35417,772 ~ $589.432 X + 268.445 X (Quadratic model Y = 36133,696 ~ 5954.738 X + 310.895 X ~ 1.347 X [Cubic model} where:Business Analytics = the forecasted or predicted total cost, and = servers. X = can be the number of com} wuadratic mode use the 4 Paterdive For purposes of illustration, we th oe Using the curve-fitting models, one can either is a minimizing value for X (number of servers) or per eee Where values of X are substituted into the model to comp! S ber ol serve! imal num ‘Asa simpler solution method to finding the opt sees teat redicted values for se predicted values ‘occurs at 9 servers at Rs. 4533, whereas t cost is Rs. 4678 occurring at 10 servers. The differences ar fi values. Like regression in general, ANOVA actual n process of curve uch the lowest 5 cost in the quai ic model demonstrate a strong rel lues, there is some error. This process provides a neat-optim® does not guarantee one. A G 1 1 30096.79 2 2 25312.69 3 3 21065.48 4 4 17355.16 5 5 14181.74 6 6 115452 a 1 9445.553 8 8 7882796 ee ee G 6856.929 6367.952 6415.865 7000.68 812.361 9780.94 15 16 1197642 1470878 17978.03 2178418 26127.21 3107.13 20 | 20 Fis, 59. Predicted total costin server problem for each server alternative Like all regression models, curve fitting is an estima the supporting statistics, ike ANOVA, provide some resulting solution m process and has risks, but legree of confidence in the 5.6. PRESCRIPTIVE ANALYTICS ANALYSIS STEP IN THE BA ANALYSIS ERAN ASS 5.6.1. CASE BACKGROUND REVIEW The predictive analytics analysis in Chapt relationship between radio and TV com future product sales. The ramifications of thes funds away from paper and POS ads to radio and TV commercials. Deter ing how much of the 350,000 budget should be allocated between the two types of Commercials requires the application of an optimization decision-making methodology. 6 revealed a statistically strong als that might be useful in predicting € results suggest a better allocation ofor Business Analytics ee The allocation problem of the budget to purchase radio and TV comme multivariable (there are two me‘ . ions on how one can allocate the budget funds),optimization problem (Ba always secks to optimize business performance). Many opt bbe employed to determine a solution to this problem. Considering the singular objective of maximizing estimated product sales, linear programming (LP) is an ideal methodology to apply in this situation. isa ined(there are some to consider). con: ization methods could 5.6.2. FORMULATION OF LP MARKETING/PLANNING MODEL In the process of exploring the al constraints on placing radio and TV commerci all the commer cal Is was set at a maximum of Rs, 350.000 for the next monthly campaign, To receive the price discount on radio commercials, a minimum budget investment in radio of Rs. 15,000 is required, and to receive the price discount on Ty commercials, a mini-mum of Rs, 75,000 is necessary. Because the radio and TY stations are owned by the same corporation, there is an agreement that for every dollar of radio commercials required, the el must purchase Rs. 2 in TV mnship found in the cation decision as an previous predictive analysis, one can formulate the budget al LP model using a five-step LP formulation procedure. Step 1. Determine the type of problem This problem seeks to maximize dollar product sales by determining how to allocate budget dollars over radio and TV commercials. For each dollar of radio ‘commercials estimated with the regression model, Rs. 275.691 will be received, and for each dollar of TV commercials, Rs. 48.341 I be received. Those two Parameters are the product sales values to maximize, Therefore, I be a prescriptive Analytics campaign X = the number of dollars to invest in TV. or campaign step 3. Formulate the objective function Because the mul le regression model defi funet ines the dollar sales i with the two independent variables, the same Sc dollar coefficients model can be used as the contribu from the regression ; n coefficients in the obj jective func is results in the following LP model objective function: weet Pi Maximize: 2 =275.691 X, + 48.341 x, Step 4. Formulate the constraints Given the information on the limitations i constraints: in this problem, there are four Constraint 1 - No more than Rs. 350,000 k lowed for the total bud; allocate 19 both rdio(X) and TV 00 coum. So add X +X and ge budget to or equal to 350,000 to formulate the first constraint as follows X+X < 350000 Constraint 2 - To get a discount on radio (X) commer 15,000 must be allocated to radio. The constraint for this X > 15000 minimum ofRs, n follows: Constraint 3 ~ Similar to Constraint 2, 19 get a discount on TV (X) com- Tracy minimum of Rs. 7,000 must be allocated to TV, The constraint for this limitation follows: X > 75000 Constraint 4 - This is a blending problem constrai the relationship as follows: Xy van, any Ne int. What is needed is to expressEy Business Analyt, which is to say, for each one uni acquire two units of X sug differently, the ts of X. Given the expression, use algebra to cross 2X= Convert 10 an acceptable constrain variables on the left side as follows: 2X-K =0 a constant on the right side and the Step 5. State the Non negativity and Given Requirements two va model is ex-pressed as follows: XX 20 Becauise these variables are in dollars. can be any real or cardinal number.) TI here: formulation of an 10 be integer values. (The te LP model formula Maximize: Z = 275.691 X, + 48.341 Xz Subjectto: -X, +X, < 350000 xX, = 15000 X, 2 75000 2X,-X, = 0 and XX, 20 5.6.3. SOLUTION FOR THE LP MARKETING/PLANNING MODEL ‘One knows that both E: and solve the bud; ing case study problem, F ‘purposes of brevity, discussion will be limited to just LINGO. LINGO is a mathematical program: language and software system. It allo the fairly simple statement of the LP model to be entered into a single window and run to generate LP solutions, 4 blank window for entering whatever ‘ype of model is Max = 275.691 + X1 + 2 <= 350000; X1 >= 15000; X2 >= 75000; 2 * x1 - x2 = 0; XL + 48.341 + x2; case study es in the model entry requirements over usual LP model fo i re LINGO. These incl ‘and a space betwee functions whe semicolon, and ( each expression with a ich aren’t necessary. Having entered t the bar at the top 350,000 promotion bud-get is 10 and Rs, 233,33330 to TV The resulting Z value, which in this model is the total predicted product sales in dol 434435208. of Rs. 43,443,524 Comparing that future ated month's product sales with the average current monthly product sales of Rs, 16,717.20 it does appear thatthe firm in this case allocate Rs, commer Oe = iaBusiness Amalie, L___.—__ === ere future estimated monthly product sales Se 3. Click on Downk study will optimally maxi »wnload a Trial Version. 4, the budget accordin, 5 0.43443822408 6. The system may r a mstem may request tha you riser your copy. Feel fee to doit now ipeseoraicon To con you downloaded it vatue correctly, variable ite 9.000000 aig 8 ae 0.000000, 5.6.4. FINAL COMMENT ON THE MARKETING/PLANNING MODEL nov stash oc Sumas Dit ‘olution methodology used to generate an allocation solution 000008 ss onases3ea ee "does not guarantee that the firm using this 2 nn 9.00 ‘echleve the results suggested in the analysis. Like any 23 . m process, the numbers are i ‘ 1se393°3 oad est Process, the numbers are only predictions, not assurances of outcomes, ‘The high levels of significance in the conformational analysis and the added use of other * (R-Square, adjusted R-Square, ANOVA, and so on) in the model development provide some assurance of predictive validity, There are many asher methods and approaches that could have been used inthis case std, Learning how fo use more statistical and decision science tools helps ensure a better solution in the final analysis. ; mm: marketing/planning case study ics analysis step brings the prior statisica ‘on-making process where a potential busines deci analytic steps into an applied mi performance improvement is shown to better this orgar pea resources more effectively. The management job of moni pee checking to see that business performance is in fact improve se in the BA analysis. Without proof that business performance is improved, ifs Unlikely that BA would continue to be used ae Jindo.com). The use of tis 30 software isa product of Lindo Systems (www. fetes a ited time as a demo for free. For purpose ‘ent, For those interested in owning a cop yn avail-able through the Lindo websi there is an inexpensive versio ce Windows @versins of LINGO are compatible with Windows 2000, Windows x Windows 7, and Windows 8. To obtain the trial version (useful fi jion’s ability to use i 5.7. DEMONSTRATING BUSINESS PERFORMANCE IMPROVEMENT THROUGH BUSINESS ANALYTICS ee 5.7.1. MEANING software can be made available for A business performance improvement (BPI) is all about understanding firm's srengths and weaknesses as an organization and what processes need to change 10 meet a short- or lon; finding the right met Microsot term business goal. Defining “performance improvement” and 5 10 measure it are the keys to success. While business process improvement is all about adjusting methodologies throughout an organization, business performance improvement focuses on measuring how these process changes can impact bottom-line financial performance and cash flow. While many organizations tum to BPI after an unexpected dip in 1, Go to www.lindo.com. 2. Click on the LINGO icon.Business Analytic, 538) ate a * financial performance, the program can also be valuable during times of stability , as.a way to increase revenue and market share. 5.7.2. THE BENEFITS OF A BPI PROJECT DONE RIGHT INCLUDE: ‘Technological improvements across organi + A better understanding of significant pain points and opportunity areas ‘Streamlined processes and communication channels Improved financial accounting and reporting 5.7.3. STEPS TO DEMONSTRATE THE BUSINESS PERFORMANCE DEVELOPMENT 2. Define the Present and Future: To improve performance, organi fons need to understand the business's curren, performance and future potential using tools and resources such as the following, 4 SWOT analysis reveals the weaknesses and strengths of the business—as we threats and opportunities. A benchmarking analysis compares the business performance against similar businesses, Market research and trend analyses look a large-scale business data to discover how to better serve the customers. 2. Pick a Few Well-Defined Goals Usually, a business with more than five major goals won't be able to focus enough resources on any one of them to accomplish it well. To make your busin. more efficient, pick only three to five goals. Make them specific, measura) achievable, relevant (improving your business), and attached to a deadline. Fi put them in order of priority, because it’s efficient to focus on some goals bet others, rather than trying to do of them at once. Some are more urgent, and some can only be achieved with a methodical, long-term plan. 3. Create a Plan To move toward achieving the goals, entity need to create a plan composed of following clements. First, list the necessary actions and tasks in detail, Ne: Jength of time for each, with start and end dates. Next, responsible for the completion of each task. Also, resources needed including budget. Fi how it will be measured , define the desired result of each task and prescriptive Analytics 44, Get the Right People on the Job better equipped to improve overall business performance ‘5, Monitor Activities and Results Finally, Ane tency in order to sustain in the market. To monitor he Progress and results of tasks using the measurement citeia you defined inthe Sep. Some simple tasks may be measured as simply “completed” or sway to each this concrete feedback, you can scarch for ways to increase the business's ef at tuming resources into outputs. ‘5.7.4. CASE STUDY: PROBLEM BACKGROUND AND DATA A Sri Kr mercial manufacturing firm is facing a supply chain problem. ‘The manufacturer produces and sells a single product, a general-purpose small motor 8s & component part to different customers who incorporate the motor into their Various finished products The manufa located in T rer has a supply chain network that connects production centers radu, Kamataka, Kerala, and Andhra Pradesh with six warehouse ies that serve commercial customers located in Chennai, Vellore, Cochin, Bangalore, Mysore, and Hyderabad. problem is the need to keep the cost of ship-ping motors The manufacturer adopted a lean management what it produces with what is demanded at each implementing this ph inability to for the customer demand month to month. If the forecast ofwi 5.40) Business Anal criptive Analytics customer demand is too low and not enough inventory is available (an underage of .e manufacturer has to rush ordes motors that end up being c the the forecast is too high and the manufacturer produces and ships unwanted inventory (an overage of inventory), the warehouse incurs wasteful storage costs, ‘manufacture ‘The management of the manufacturing firm has decided that an analytics-based procedure needs to be developed to improve overall business performance. This would be a procedure that analysts could use each month to develop an optima A key part of this procedure would be to include a means to accurately forecast customer demand and an optimization process for ipping products from the ‘manufacturing centers to the warehouse demand destis ‘The manufacturing firm created a small BA team to develop the procedure. The BA team consists of a BA analyst(who would be responsible for using the proced and heads the BA team), the supply chain gener manager, the shipping mana (responsible for drafling the shipping schedule), and a warehouse manager(whose to develop monthly forecasts. Determining a procedure by which analyst teams can determine oj shipments between supply sources and demand destinations re-quires differing ty of data. There is supply, demand, and cost data required to plan shipments. The manufactured supply of motors produced at the Tamilnadu and Kerala pl determined once the forecast demand is established. The BA team established there is ample capacity between both plants to satisfy the forecasted customer demand atthe six warchouse demand desti tions. The BA team determined that the cost data for shipping a motor fror production centers to the customers depends largely on distance be-tween the where the items are trucked directly by the manufacturer to the, warehouses. The cost data per motor shipped to a customer is given in Table 5.3. For exampl ‘manufacturer % 4 per motor to ship from Tamilnadu to Karnataka. These cost are routinely computed by the manufacturer's cost accounting department and are assumed by the BA team to be accurate. sal Supply | Chennai, | Vellore | Cochin, | Bangalore, | Mysore, | Hyderabad Center [ Tamilnadu Karnataka) R=4 | Rs6 | Rs9 | Roe Rs.5 Rs.6 ‘Andhra Rs5 | Rs8 | Rs2 | Rss Rs.8 Rs. 5 Pradesh : Table 5.3. Estimated Shipping Costs per Motor The present system of forecasting customer demand usualy results in costly overages and under ages shipped to the warehouses. In the past, the manufacturer would take a three value smoothing average to estimate the monthly demand. This evolved by taking the last three months of actual customer motor demand and averaging them to produce a forecast for the next month. The process was repeated each month for each of the six warehouses. Not making products available when customers demanded them caused lost sales, so the manufacturer would rush and ship Products to customers at a loss. On the other hand, producing too much inventory meant needless produ: inventory, and shipping costs, To deal with the variability in customer demand forecasting, models for each Warehouse's customer demand would need to be developed. The customer demand data on which to build the models was collected from prior monthly demand in ‘motors. To determine which data to ude in a final sample and which to exclude, a ly useless and out-of- Going back more than 27 months invited eyclical variation$ caused by change: economy that were no longer present, so that data was removed. Unfortunat few simple rules were adopted to el data collected on warehouse cus-tomer monthly demand files is presented in Table 34. It was decided that the most recent three months (darkened months of 25, 26, and 27) would not be included in the mode! development, but instead would be used for ‘ion purposes to confirm the forecast accuracy of the resulting models, This lar to what was referred to s a training data set and aBusiness Analytics Prescriptive Analytics 543 DESCRPIVEANALTICS Natya Month | Chennai| Vellore | Cochin | Bangalore | Mysore | Hyderabad DESCI 1 3120) 2130 3945 14020] 5045 4610 ics Table 2 3090| 2290] 4000 13890} 5030 4630 Vellore _| Cochin | Bangalore | Mysore | Hyderabad 3 3140| 2405] 410s 13785) 5075 4650 3010-8696 | 3186.73913 | 4521.95652| 9904.565 | s020.652| 478622 75} S015 4680 4 3010) 2580) 4300) 135 11.993296 | 112.691475 | 69.635371| 708.1243 | 3.241046] 2ss1144 5 2900| 2635| 4255 13345] ois 4700 6 2990| 2690| 4420 12990} 5020 4750 3000 3120 4535 8880| ois 475 7 3000] 2740 4540 12340] 5025 4800 ue 0 8 3030| 2780 4670 11850) soso 4865 Mode’ 2990 NIK aro BNIA S015 4680 9 3050] 2890] 4820 11010} so10 4910 Standard | 57.517827 | $40.449326 | 333.959507 3396.045 | 1554351 | 122-3486 10 2970) 2940) 4780 10015] so10 4980 | Dev. u 2980! 3000 4900, 9875 | S015 5000 Sample 3308.3004 | 292085.474 | 111528.953 | 11533120 | 241.6008] 14969.17 2 2990| 3020) so20 9005) sois S010 Var. a BZ 22 la 2045) 2880)|gegSCLO A220, Kunosis | 0.4632145 | -1.1215079 =| 1.45505 | 6.482162] 1.04279 14 3100} 3180] 4o4s 7990| sos 4900 12708889 1S a me 210) i 855 Bie et po Skewedness | 0.1407847 | 020384211 | -0.0337087 0.142149 | 2.442156 | 0.571352 16 3000} 3270] 4780 6920] 5020 4800 Range 240 1740 104s] 9860 65 380 7 3040] 3455} 4650 6745| soio 4785 18 3060) 3575) ass 6010| sois 740) Miniinum, 2900 2290 4000] 4030}. so10 4630 19 2950) 3765] - 447s 670| so4o 4700 | Maximum 3140 4030 5045] 13890] so7s 5010 20 2970) 3810) 4330 $345| so10 4695 Sum 69250] 73295] 10400 204805] 115475] 110090 2 2990} 3910] 4325, 5110] so20 4690 Count 2 23 2 23 2B 2B 2 3060| 3990) a1ss 4760. i au mile Largesi(1) 3140 4030 504s} 13890] 07s so10 2B 3000) 4010} 4090 4320 sois 4670 Pe 4 4 3010 4030] 4010 4030 so10 4660 re Per pec ael ta la 4030 we 1, 25 2980 4285} 3720 3005} soio 4590 Confidence | 24.872573 | 233.707814| 144.41492| 1468.56 | 6.721519] s2.90748 6 2965| 4420|. 3530 2515}. “soio 4570 Meal mea 2945} 4560| 3330 2030| soos 4555 ee) Table SA. Actual Monthly Customer Demand in Motors Table 5.5. Excel Summary Statistics of Actual Monthly Customer Demand in Motors pe