ANALYTICS FOR ENABLING BUSINESS STRATEGY PROFESSOR CATHAL BRUGHA, M.B.A, PH.D., FMII FOUNDER DIRECTOR OF THE ANALYTICS INSTITUTE PRESIDENT OF THE ANALYTICS SOCIETY OF IRELAND WHO DO ANALYTICS? TECHI-ANALYSTS DISTINCT: PARTNERS IN THE PRACTICE OF ANALYTICS Assessing the Viability of using Open Source Tools for Marketing Analytics – Niamh Carroll and Paul Jones, 2008 -2009 Decision Support System for Credit-Scoring – Michael Wilson and Simon Shortt: 2009 -2010 Application of Network Analysis in Insurance Data Sets in order to highlight potentially Fraudulent Claims Cliona Fleming and Colman Horgan, 2010-2011 Improving the Accuracy and Efficiency of Predictive Modelling in Classification and Regression Trees – Eoin Fitzpatrick and Ciarán Tobin: 2011-2012 Types of Analytics Quant Approach Applications Qual Approach DECISION ANALYTICS : THEORY BUILT ON PRACTICE Development of a Muti-Criteria Decision Support System for Early Diagnosis of Dementia in the Elderly A Multi-Criteria Approach to the Construction of Team Performance Indicators in Professional Rugby Union Cultural Comparisons Between China and Ireland Reform of the Culture in the Public Service DEVELOPMENT OF A MUTI-CRITERIA DECISION SUPPORT SYSTEM FOR EARLY DIAGNOSIS OF DEMENTIA IN THE ELDERLY Apostolos Tsakmakis and Muhammad K Hafeez Dr. Radu Marinescu and Dr. Léa A Deleris IBM Ireland Research Laboratory Professor Mary McCarron – Dean of the Faculty of Health Sciences, Dr Kate Irving - Nursing and Human Sciences Dr Allys Guerandel - Integrating E-learning with Psychiatry teaching, Dr. Aurelia Ciblis – UCD School of Medicine and Medical Science Dr. Abdul Rauf - GP in Kilkenny (Software Test) WHAT IS DEMENTIA? Definition: Dementia is the term used to describe a collection of symptoms caused due to the loss of cognition, behavioural changes and a decline in social activities in the elderly Dementia of Alzheimer’s type (AD) Mostly in Elderly people age > 60 More than 50% of Dementia in elderly is of AD 41,000 dementia patients and number will be more than 100,000 by 2036 DIAGNOSIS OF DEMENTIA American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) The World Health Organization's International Classification of Diseases (ICD) IMPORTANCE OF EARLY DIAGNOSIS Misdiagnosed/late diagnosis by GPs (35% responsibility) Disease can be Managed Improved Quality of Life Less stress for Family members Fewer Patients requiring specialized help Fewer referrals to Memory Clinics Symptoms (Criteria) for Dementia PSYCHOMOTOR ACTIVITIES Memory Impairment Sleep Disturbances Language disturbance Patient’s Medical History Personal Care Orientation Attention/Thinking/Conciousness Social Activities/hobbies Physical Exam Mood/Behaviour Instrumental Activities of daily living MOST COMMON DISEASES WITH THE SAME SYMPTOMS Depression Delirium DEMENTIA DIAGNOSIS A MULTI-CRITERIA DECISIONMAKING (MCDM) PROBLEM Multiple Diagnostic Criteria Conflicting Alternatives STRUCTURING THE CRITERIA TREE Self Self Others World Physical Psychological Social Psychomotor activities Others Physical Exam Medical History Memory Personal care Impairment IADL Language Mood/Behaviour Social Activities Disturbance World Sleep patterns Orientation Attention/thinking/ consciousness IADL: Instrumental Activities of Daily Living Convince me they have Dementia: how they relate to Self, Others, the World (structure) DIAGNOSTIC GUIDELINES Depression Criteria Delirium (DEL) Dementia (DEM) (DEP) Mild Moderate Severe Noticeable Hypoactive disturbance Large No disturbance (retardation) in movement. in movement Movement Retardation: Decreased More Psychomotor Prolonged Activities restlessness, Slightly Hyperactive little risk no control in behaviour, psychomotor considerable activity, intense highly reactions intense restless, occasional (agitation) little risk behaviour reactions motivation Apraxia: Psychomotor changes characteristically Agitation: agitated depression occurring late in the illness CRITERIA WEIGHTS AND SCORING OF ALTERNATIVES Imprecise Weights User is not able to provide a specific importance to one symptom over another. Verbal Scale Direct-interactive Structured-Criteria (DISC ) System Utility Scoring (DISCUS) System Relative Intensity Measurement (DISCRIM) System DECISION SUPPORT SYSTEM (DSS) REQUIREMENTS User friendly interface Efficient scoring process Accessibility to historical data Data protection Flexibility Integration with other DSS MCDM IMPLEMENTATION SOFTWARE Imprecise input requirements Additional software requirements (EXCEL) User friendliness Generic approach PROTOTYPE SOFTWARE Stand-alone Java based application Software Implementation for MCDM model Handles Imprecise Inputs Data Protection Provides functionality for further analysis CONTRIBUTION TO BUSINESS & SOCIETY Early Diagnosis can reduce the cost of Dementia care (approx €442/person in late stages). The quality of life of the Person with Dementia can be improved through therapies in the early stages. Reduced number of misdiagnosed patient referrals to Memory Clinics Improved quality of life for Dementia patients Correct early clinical diagnosis with no Dementia can reduce extra cost (MRI, Brain scans etc.) ACADEMIC CONTRIBUTION Structuring DSM IV using MCDM Imprecise inputs for MCDM model Structured criteria combined with diagnostic guidelines matrix can be used for improvement of dementia training for GPs A MULTI-CRITERIA APPROACH TO THE CONSTRUCTION OF TEAM PERFORMANCE INDICATORS IN PROFESSIONAL RUGBY UNION Alan Freeman and Declan Treanor Opta Sports CURRENT RUGBY UNION & TEAM PERFORMANCE METRICS • • Current methods used to objectively depict team performance leave it up to the expert user to make sense of them, using technical and qualitative analysis High Level information flow looks like this: Match Event Data Performance Metrics Expert User Input Post Match Analysis QUESTION: RUGBY UNION & TEAM PERFORMANCE METRICS • Can we use Business Analytics to introduce expert user knowledge earlier in the process so as to produce a team metric that offers a more meaningful description of performance? Match Event Data Performance Metrics Expert User Input Post Match Analysis INTRODUCTION: RUGBY UNION & TEAM PERFORMANCE METRICS • Can we use Business Analytics to introduce expert user knowledge earlier in the process so as to produce a team metric that offers a more meaningful description of performance? Match Event Data Expert User Input Performance Metrics Post Match Analysis RUGBY UNION & TEAM PERFORMANCE METRICS • • • • • Hot Performance Indicators constructed using existing Multi Criteria Decision Making tools Expert Users brought through steps so as to be convinced of good performance. Actions need to be executed well technically in any given context (depending on opposition) and improve the team’s situation , e.g. gives some advantage – field position or on the score board Can be used to analyse comparative team performance by highlighting imbalances within underlying adjusting structure Also, provide a basis for match outcome prediction, using a simple Time Series Forecasting method RESEARCH QUESTION & SUCCESS CRITERIA • • • Can the factors that contribute to team performance in Rugby Union be considered to follow an underlying adjusting structure? Using this underlying structure, can a new team performance metric be created that will lend itself well to comparative analysis of teams and match outcome prediction? Success = Successful proof of concept • Derived initial criteria relating to possession, set-pieces, distribution, general execution (Hughes and Bartlett, 2002) • Criteria: possession, set-pieces, distribution, general execution Came from soccer – experts extended them to eight for rugby • We consider the sub-criteria for each of the eight criteria, for example Tackles • • • • Sought to convince expert users by looking at different aspects. Ensure technique, appropriate to opposition and gives team a game advantage Example for Execution>>Tackles Able to identify relevant Match Events for construction of HPI metrics Should be convincing: Technically in the particular Context in the actual Situation • Base Scores • • 83 match event outcomes given scores, positive & negative outcomes Score Modifiers Positional clusters (James et al. (2005)) • Quality of Opposition ((Taylor et al., 2008)) • Location on Pitch • Home Advantage / Away Disadvantage (Nevill et al., 2007)) • • Base Scores • • 83 match event outcomes given scores (positive & negative outcomes Score Modifiers Positional clusters (James et al. (2005)) • Quality of Opposition ((Taylor et al., 2008)) • Location on Pitch • Home Advantage / Away Disadvantage (Nevill et al., 2007)) • METHODOLOGY: DATA & SOFTWARE USED • • • • • Data Provided by Opta Sports 893 Extensible Markup Language (XML) Files (2008 – 2011 data) Only used data between 2009 – 2011 – data quality issues XML included fixture data, player data and match event data Software used: MySQL, Java Snippet of XML provided ENTITY RELATIONSHIP DIAGRAM BUSINESS CONTRIBUTION • • • • Metric shows difference in performance / form over time. Example, Celtic League 2011 Munster, Leinster finished top, Aironi finished bottom. See HPI over season Can be used to highlight strengths / weaknesses viz a viz opposition teams BUSINESS CONTRIBUTION • • • • Well structured teams should show balance among the factors contributing to their performance Methodology highlights imbalance Metrics form basis for prediction of match outcomes (win or loss) Useful for coaches, sports management, bookmakers Actual HPI vs Match Outcome SUCCESS CRITERIA • Can a new team performance metric be created that will lend itself well to comparative analysis of teams and match outcome prediction? View the Imbalance – Top vs Bottom of League Competition # Fixtures # Correct % Correct Heineken 2010 75 54 72.0% Heineken 2011 78 60 76.9% Heineken 2012 72 53 73.6% Magners 2010 81 50 61.7% Magners 2011 134 106 79.1% Rabo 2012 112 73 65.2% Overall 552 396 71.7% Predicted HPI vs Match Outcome Competition # Fixtures # Correct % Correct Heineken 2010 63 44 69.8% Heineken 2011 66 38 57.6% Heineken 2012 60 38 63.3% Magners 2010 75 41 54.7% Magners 2011 128 80 62.5% Rabo 2012 106 66 62.3% Overall 498 307 61.6% • • • • • Academic Contribution – Novel application of MCDM Business Contribution – new ways to analyse comparative performance and predict future performance Learning: How to deal with Professionals / Organisation outside our normal comfort zone (usually IT / Finance). Successful in terms of research questions. Married quantitative with qualitative approach Further research • Refinement of scores and modifiers • Improve scoring methodology (e.g. team rankings, referees) • Expand scope – better forecasting method (e.g. Artificial Intelligence BUSINESS CONTRIBUTION • • • • Metric shows difference in performance / form over time. Example, Celtic League 2011 Munster, Leinster finished top, Aironi finished bottom. See HPI over season Can be used to highlight strengths / weakness viz a viz opposition teams ACADEMIC CONTRIBUTION • • • • Novel application of business analytics to a real world problem Viewed sport as an Adjusting Process and showed link to existing structures Practise based approach to evaluating team performance Usable in other contexts (different sports, different decision problems) SUBJECTIVE COMPARISONS BETWEEN CHINA AND IRELAND ©C ATHAL Introverted Development Committing Phases Extroverted Development - Convincing Stages Tech – Self India - Ireland x x x x Somatic – Need – Thinking – Fear Analysis – Abduce Psychic – Prefer – Feeling – Anxiety Design – Deduce Pneumatic – Value – Knowing – Resent Implement – Adduce M BRUGHA 1. xx x x xx Physical / Intuiting / Affection Contextual – Others - China x x x x xx x x xx Situational – World - U.S. x x x x x xx xx x 2. Political / Recognizing / Comradeship 3. Economic / Believing / Partnership 6. Emotional / Trusting / Empathy Finding 5. Cultural / Learning / Friendship Filling 4. Social / Sensing / Sexual Fitting 7. Artistic / Experiencing / Collaboration 8. Religious / Understanding / Communion 9. Mystical / Realising / Charity OBJECTIVE COMPARISONS BETWEEN CHINA AND IRELAND ©C ATHAL M BRUGHA OBJECTIVE COMPARISONS BETWEEN CHINA AND IRELAND ©C ATHAL M BRUGHA OBJECTIVE COMPARISONS BETWEEN CHINA AND IRELAND ©C ATHAL M BRUGHA Introverted Development Committing Phases Reforming the Culture in the Public Service, which came from 19th Century Britain that protected corporate entities, and feared individuals and representative groups Extroverted Development - Convincing Stages Individual Self - Technical Representative Corporate – Others - Contextual World - Situational No individuals with given roles other than ministers and some ombudsmen, Governor of Central Bank, regulators, etc. Some external boards: c. 400 state authorities, little internal lateral or mainly unconnected, de vertical coordination. facto independent. Little oversight of Can postpone, ignore, authorities. Councils little sideline suggestions, power over officials pass the “hot potato” Somatic – Need – Thinking – Fear Analysis - Abduce 1. Psychic – Prefer – Feeling – Anxiety Design – Deduce Pneumatic – Value – Knowing – Resent Implement – Adduce Physical / Intuiting / Survey 2. Political / Recognizing / Study 3. Economic / Believing / Define 6. Emotional / Trusting / Acquire Finding 5. Cultural / Learning / Design Filling 4. Social / Sensing / Select Fitting 7. Artistic / Experiencing / Construct 8. Religious / Understanding / Deliver 9. Mystical / Realising / Maintain Introverted Development Committing Phases Reforming the Culture in the Public Service, which came from 19th Century Britain that protected corporate entities, and feared individuals and representative groups Extroverted Development - Convincing Stages Individual Self - Technical Representative Corporate – Others - Contextual World - Situational No individuals with given roles other than ministers and some ombudsmen, Governor of Central Bank, regulators, etc. Some external boards: c. 400 state authorities, little internal lateral or mainly unconnected, de vertical coordination. facto independent. Little oversight of Can postpone, ignore, authorities. Councils little sideline suggestions, power over officials pass the “hot potato” Somatic – Need – Thinking – Fear Analysis - Abduce 1. Psychic – Prefer – Feeling – Anxiety Design – Deduce Pneumatic – Value – Knowing – Resent Implement – Adduce Physical / Intuiting / Survey 2. Political / Recognizing / Study 3. Economic / Believing / Define 6. Emotional / Trusting / Acquire Finding 5. Cultural / Learning / Design Filling 4. Social / Sensing / Select Fitting 7. Artistic / Experiencing / Construct 8. Religious / Understanding / Deliver 9. Mystical / Realising / Maintain Reform of the Culture in the Public Service ©Cathal M Brugha ANALYTICS FOR ENABLING BUSINESS STRATEGY PROFESSOR CATHAL BRUGHA, M.B.A, PH.D., FMII FOUNDER DIRECTOR OF THE ANALYTICS INSTITUTE PRESIDENT OF THE ANALYTICS SOCIETY OF IRELAND WHO DO ANALYTICS? TECHI-ANALYSTS ANALYTICS HAS MANY APPLICATIONS COMBINED SUBJECTIVE / OBJECTIVE COMPARISONS ©C ATHAL M BRUGHA