Bookflare net SAS for finance

SAS for Finance Forecasting and data analysis techniques with real-world examples to build powerful financial models Harish Gulati BIRMINGHAM - MUMBAI SAS for Finance Copyright © 2018 Packt Publishing All rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, without the prior written permission of the publisher, except in the case of brief quotations embedded in critical articles or reviews Every effort has been made in the preparation of this book to ensure the accuracy of the information presented However, the information contained in this book is sold without warranty, either express or implied Neither the author, nor Packt Publishing or its dealers and distributors, will be held liable for any damages caused or alleged to have been caused directly or indirectly by this book Packt Publishing has endeavored to provide trademark information about all of the companies and products mentioned in this book by the appropriate use of capitals However, Packt Publishing cannot guarantee the accuracy of this information Commissioning Editor: Amey Varangaonkar Acquisition Editor: Divya Poojari Content Development Editor: Amrita Noronha Technical Editor: Nilesh Sawakhande Copy Editor: Safis Editing Project Coordinator: Shweta H Birwatkar Proofreader: Safis Editing Indexer: Aishwarya Gangawane Graphics: Jisha Chirayil Production Coordinator: Shantanu Zagade First published: M ay 2018 Production reference: 1250518 Published by Packt Publishing Ltd Livery Place 35 Livery Street Birmingham B3 2PB, UK ISBN 978-1-78862-456-5 www.packtpub.com mapt.io Mapt is an online digital library that gives you full access to over 5,000 books and videos, as well as industry leading tools to help you plan your personal development and advance your career For more information, please visit our website Why subscribe? Spend less time learning and more time coding with practical eBooks and Videos from over 4,000 industry professionals Improve your learning with Skill Plans built especially for you Get a free eBook or video every month Mapt is fully searchable Copy and paste, print, and bookmark content PacktPub.com Did you know that Packt offers eBook versions of every book published, with PDF and ePub files available? You can upgrade to the eBook version at www.PacktPub.com and as a print book customer, you are entitled to a discount on the eBook copy Get in touch with us at service@packtpub.com for more details At www.PacktPub.com, you can also read a collection of free technical articles, sign up for a range of free newsletters, and receive exclusive discounts and offers on Packt books and eBooks Contributors About the author Harish Gulati is a consultant, analyst, modeler, and trainer based in London He has 15 years' financial, consulting, and project management experience with leading banks, management consultancies, and media hubs He enjoys demystifying his complex line of work in his spare time This has led to him being an author and orator at analytical forums He has also co-authored Role of a Data Analyst, published by the British Chartered Institute of IT (BCS) He has an MBA in brand communications and a degree in psychology and statistics Clustering at some level is a business decision that isn't based on a purely statistical basis The overall aim of segmentation should be to find clusters of customers that have a statistical rationale, but also to segment the customers into a manageable number of clusters that help produce effective strategies Prof Cox decided to go for the five-cluster solution and see what impact it had on the segments generated Unfortunately, the SAS University Edition on which the modeling has been done for the book doesn't support some visual features that help to effectively showcase the output of Proc tree However, the full edition of SAS software produces a more visually appealing chart of the procedure For our current consumption, only the tabular output has been shared in Figure 7.17 The procedure is going to help us to assign the cluster number to each of our customer IDs This is an important step, as after this we can go ahead and try to understand the profiling of our cluster constituents This is a Proc tree five-cluster specification: Proc tree data = tree out = cluster_output nclusters=5; Id custid; Copy age aum risk_appetite fund_performance investment_potential investment_involvement complex_product; Run; proc print data=cluster_output(drop=clusname); run; Figure 7.17: Customer cluster allocation Although there was a fair bit of confidence in going with the model generated up to now, there were a lot of other alternate models considered in the build phase The models differed in the type of clustering used, the data standardization used, and the mix of variables used to build the models Let's produce an alternative model using the following code: /*Age and AUM have been dropped in the model*/ Proc varclus data=cluster_model; Var risk_appetite fund_performance investment_potential run; investment_involvement complex_product; The output in Figure 7.18 shows that the model did not produce any splitting and suggested a singlecluster solution This happened after two key variables were dropped from the modeling solution: The chart reconfirms that we are left with a single cluster solution after the omission of two variables from the alternate model code: Figure 7.18: Alternate modeling attempt Prof Cox gave the five clusters a name and recommended a strategic direction for Vogue to take regarding the clusters As a modeler, the business expects that the model built is statistically robust It can be validated and is documented with a high degree of governance during the build, approval, and implementation phases However, modelers at times tend to forget how simple summaries of the model can help to create a greater understanding of the model and help trust the insights generated The next few pages are dedicated to showcasing how Prof Cox summarized the output from the model: Figure 7.19: Segment profile as defined by the modeler She further described the segments in Figure 7.19 as: Star performers: These customers are the ideal age group Probably the happiest Vogue customers in terms of returns on investment They have a high level of AUM and there is a potential for further investment They are medium risk takers and have no preference regarding simple or complex products They are the smallest cluster of the five Cash cows: They are split across the age groups They behave like cash cows for Vogue as they have low-to-medium AUM, yet are large in number and help to sustain the business They are involved in their decisions and aren't risk takers 76% of them have experienced low returns with Vogue However, due to their low risk-taking preferences, and in general, lack of inertia to move to competitors, they are thought to be good for the business in the long run 18% of them have experienced medium returns and this percentage could be further increased in the years to come Nurture: They are split across age the groups and have a medium-to-high-potential to invest They have experienced a mixed level of fund performance Their risk appetite is quite high and they don't tend to have complex products The most promising feature of this segment is that they have low-to-mid-level AUM Given their potential to invest, with a bit more focus this segment could become a more exciting segment for Vogue After all, this is their second biggest segment Keen but not there: A fairly big cluster of younger individuals with lower involvement, low fund performance, low potential to invest, and low-to-mid-AUM These are high risk takers who tend to have complex products They could also benefit from being nurtured, but they are probably not mature enough from a prospective client perspective This is a segment that Vogue should continue to watch out for and be patient with Going nowhere: This segment has young members They have low involvement in investing and lower potential The lower potential might stem from the fact that they already tend to have high AUM with Vogue However, the predominantly low fund performance that they might be facing could be a factor in them expressing lower investment potential Vogue needs to look at this segment and see if relationship managers need to focus less on these customers So, how did Prof Cox define and describe the segments? The output dataset produced as part of the proc tree code in Figure 7.17 is where the cluster name is stored against each customer However, the cluster name is a number and isn't a descriptive field describing the characteristics of the cluster/segment Prof Cox had to produce some profiling tables from the output dataset Let's look at them: Age No of Customers Cluster Young Mid Senior 14 22 2 36 36 24 30 18 42 35 10 24 AUM No of customers Cluster Low Med Good High 31 25 59 12 30 33 28 20 Risk appetite No of customers 35 10 Cluster Low Med High 32 70 25 13 53 11 37 24 13 Fund performance No of customers Cluster Low Med High 1 29 73 18 20 13 39 39 33 12 Investment potential No of customers Cluster Low Med High 1 35 84 12 30 27 21 42 Investment involvement Cluster % of customers 13% 40 32% 24% 16% 15% Complex Product No of customers Cluster No Yes 19 19 22 74 55 17 42 38 Cluster % of customers 13% 32% 24% 16% 15% Cluster summary Cluster Age AUM Risk appetite Fund performance Young to medium High Medium High Split across Low to medium Low Low Split across Low to medium High Split across Young Low to medium High Low Young High Split Across Low Cluster summary continued: Cluster Investment potential Investment involvement Complex product High Medium to high No preference Low High Yes Medium to high Medium to high No Low to medium Low Yes Low Low Yes Cluster Segment Star performers Cash cows Nurture Keen but not there Going nowhere Figure 7.20: Segment allocation details As you can see in Figure 7.20, we need to develop some profiling tables to understand the characteristics of our segment constituents We can already see that the clusters differ between the mix of age groups, assets under management, and risk appetite By adding profiling info using other variables for clustering, we can come up with descriptions of the segments What Prof Cox has done is to go a step beyond and name the segments Naming the segments is an easy way to remember the key characteristics of the segments Some businesses prefer to name segments in a way that explains what each segment means to their strategy There is no specific scientific way to describe and name a segment The insights from profiling should make business sense, and the naming of segments should lead to the achievement of some business goals Having named her segments and being ready with her proposed strategy for the segments, Prof Cox looked forward to sharing the modeling results with Vogue Summary We have looked at two analysis methodologies in this chapter MBA analysis and segmentation Both use datasets related to time series, but they transform the data and don't consider the time element explicitly as part of the analysis In the MBA analysis, we saw that sequence was a sort of proxy for time series We generated insights by focusing on business problems that were the subject of modeling MBA lacks the statistical depth and rigor that clustering has However, neither are strictly statistically-driven analysis scenarios With MBA, we showcased how it makes intuitive sense to evaluate the association between the products and services offered by a bank and leverage the information In segmentation via clustering, we showcased how the number of clusters generated differed from the preferred statistical information available to us from two different methodologies Yet, we decided to go for an approach that made business sense and could be supported by showcasing the analysis conducted Other Books You May Enjoy If you enjoyed this book, you may be interested in these other books by Packt: Big Data Analytics with SAS David Pope ISBN: 978-1-78829-090-6 Configure a free version of SAS in order hands-on exercises dealing with data management, analysis, and reporting Understand the basic concepts of the SAS language which consists of the data step (for data preparation) and procedures (or PROCs) for analysis Make use of the web browser based SAS Studio and iPython Jupyter Notebook interfaces for coding in the SAS, DS2, and FedSQL programming languages Understand how the DS2 programming language plays an important role in Big Data preparation and analysis using SAS Integrate and work efficiently with Big Data platforms like Hadoop, SAP HANA, and cloud foundry based systems IBM SPSS Modeler Essentials Jesus Salcedo, Keith McCormick ISBN: 978-1-78829-111-8 Understand the basics of data mining and familiarize yourself with Modeler’s visual programming interface Import data into Modeler and learn how to properly declare metadata Obtain summary statistics and audit the quality of your data Prepare data for modeling by selecting and sorting cases, identifying and removing duplicates, combining data files, and modifying and creating fields Assess simple relationships using various statistical and graphing techniques Get an overview of the different types of models available in Modeler Build a decision tree model and assess its results Score new data and export predictions Leave a review - let other readers know what you think Please share your thoughts on this book with others by leaving a review on the site that you bought it from If you purchased the book from Amazon, please leave us an honest review on this book's Amazon page This is vital so that other potential readers can see and use your unbiased opinion to make purchasing decisions, we can understand what our customers think about our products, and our authors can see your feedback on the title that they have worked with Packt to create It will only take a few minutes of your time, but is valuable to other potential customers, our authors, and Packt Thank you! .. .SAS for Finance Forecasting and data analysis techniques with real-world examples to build powerful financial models Harish Gulati BIRMINGHAM - MUMBAI SAS for Finance Copyright... of: WinRAR/7-Zip for Windows Zipeg/iZip/UnRarX for Mac 7-Zip/PeaZip for Linux The code bundle for the book is also hosted on GitHub at https://github.com/PacktPublishing /SAS- for- Financ e In case... Budget and Demand Forecasting The need for the Markov model Business problem Markovian model approach ARIMA model approach Markov method for imputation Summary Inflation Forecasting for Financial