Projects

Applied Statistics Projects

  • Enhanced the performance of general insurance pricing models using Generalized Linear Models of customer retention data, premiums, and marketing-mix data
  • Data Science, Customer Modelling, E-Commerce and Business Development in Bank Financial Management (UK and Nigerian perspectives): explores ongoing innovations in bank business models post-2007 global financial crisis
  • Modelling the growth and profitability of car insurance purchased through multiple channels
  • Systematic Stock Market Characterisation and Development (SSMCD) across key market sectors (e.g. overall market index, banking sector, oil, and telecommunications) of emerging African markets like Nigeria and Libya
  • Stochastic-Time Series Modelling for the Banking Sector of the Nigerian Stock Market (NSM) (2014-2017
  • Comparative Times-Series and Macroeconomic Modelling of Oil Price Fluctuations in Libya and Nigeria, for improving the financial performance of national upstream and downstream oil businesses in the two countries (2014-2017)
  • PhD research to develop a Digital Banking Customer Experience Enhancement Framework (DiBCEEF) in the UK using combined data-driven insights from customers, bank emplyees and senior management (2013-2017)
  • PhD research on derivative pricing in the Nigerian Stock Market
  • Statistics, business analysis, and enterprise modelling techniques on a ‘Construction Industry Statistics System (CISTATS) Performance Review’ for the Department of Trade and Industry (DTI), UK
  • PhD research to develop an Integrated Born-Global Business Modelling Template (IBMT(c)), for creating fast-growing internet-based born-global firms
  • PhD research on embedding employability skills in hospitality and tourism higher education curricula
  • PhD research on determinants of foreign direct investment in South Africa
  • PhD research on Natural Language Processing using computational learning theories, Machine Learning, language models an algorithms
  • MSc research on factors affecting economic growth in Nigeria – using time series analysis
  • Questionnaire design and analysis of data in adolescents’ use contraceptives
  • Path analysis and structural equation modelling in digital banking
  • Meta-analysis in medical research (BSc nursing)
  • Repeated measures analysis using ANOVA, MANOVA, and mixed effects modelling of educational and career performance data
  • Statistical quality control and Sigma-Sigma methodology applications in engineering
  • PhD research on comparative business modelling, qualitative and mixed data analysis in higher education
  • Statistical and geographical information systems modelling of Boko Haram terrorist data in Nigeria
  • Data transformation and regression diagnostics in motor insurance
  • Data transformation and regression diagnostics in stock market analysis, and energy price modelling
  • Data mining algorithms in car insurance: using logistic regression, decision trees, and neural networks
  • Market segmentation analysis and integrated model building for car insurance in the UK
  • A wide range of CRO/ stages I-IV clinical trial studies, see examples below
  • Biostatistics projects using generalised linear mixed effects modelling in the SAS environment with appropriate high-resolution plots
  • Investment and financial risk modelling in insurance, banking and business
  • Qualitative data analysis using NVivo software in SME and born-global firms business development research
  • BSc Nursing degree project review and proof-reading for an Open University UK Professional Practice in Adult Nursing module
  • BSc Applied Computing project review and proof-reading for Customer Relationship Management of Fast Food Companies
  • MBA HRM study skills, coursework writing, dissertation review and proof-reading for Challenges in SME Internationalisation of Nigerian Manufacturing Firms
  • MSc Business research design and data analysis for Environmental Management – case study and questionnaire design
  • MSc Applied Statistics personal tutorials in Linear Regression, Statistical Computing and Bayesian Inference for a student from the University of Leeds UK
  • MSc SPSS analysis of biomedical data, key biostatistics analyses comparing placebo versus treatment effects
  • MSc Banking and Finance project support for Determinants of Dividend Policy for Selected Firms in the Nigerian Stock Market’
  • MSc Applied Statistics research supervision for Modelling the Volatility of Share Prices for a Nigerian Bank – using ARCH-GARCH volatility analysis techniques in EViews
  • MSc Banking and Finance research support and dissertation review for An Empirical Analysis of Credit Risk Management and Bank Profitability in Nigeria – using regression models in SPSS and Excel Analysis
  • MSc Banking and Finance research support and dissertation review for The Impact of Financial Development of Economic Growth in Nigeria – using regression models in SPSS and Excel Analysis
  • MSc Public Health data analysis and dissertation review for Understanding and Interpreting the Determinants of Condom Use Among Adolescents in Selected Slums in Ghana – Questionnaire Design, SPSS and Excel Data Analysis
  • MSc Applied Computing research support for systematic critical review and dissertation plan for Creating a Computing Laboratory and Library Resources for Teaching Internet of Things in Universities – Experience from Raspberry Pi software
  • MSc Banking and Finance research support, dissertation review and paper publication guides for Entrepreneurial Finance Models for a Small-Scale Online Nigerian Enterprise – using qualitative analysis in NVivo
  • MSc Information Security Management coursework writing support, project design and review
  • PhD Knowledge Management proposal writing and admissions support for a Corporate Academic Model-Based Career Self-Management and Performance Model – using the 3E Research Methods Canvas©
  • PhD Hospitality and Tourism Education dissertation research design, data analysis and publishing guides for Emerging Trends and Influences in Ghanaian Hospitality Industry, Education and Employability – focus group and stakeholder interview guides, questionnaire design using Item Response Research Question mappings, mixed methods data analysis, thesis proof-reading and publishing guides
  • PhD Mathematical Finance and Investing proposal writing support for Uncertainty and Random Matrix Theories for Portfolio Optimisation and Risk Management, Pan African University Institute of Science & Technology, Jomo Kenyatta University, Kenya
  • PhD Mathematical Modelling of Epidemics support for admissions, University of Sheffield, UK
  • So many other projects … see more examples in the Selected Publications section.

Specialist Biostatistics/CRO and Health Economic Modelling Projects

Data Mining Techniques and Breast Cancer Prediction: A Case Study of Libya 2012: Libyan Government)

  • Used advanced data mining tools in R to implement separate and hybrid logistic regression, decision trees and neural network modelling of cancer data from Libya.
  • Compared the relative performances of the models for predicting the incidence of cancer and discovered the best performing predictive models mainly amongst the hybrid models

Several graduate and contract research projects in CRO and life-sciences contexts

  • Providing confidential contract research support to independent CRO analysts on advanced statistical procedures;
  • Skills used include Data management, SAS programming, Biostatistics, Design of Experiments, Health Economics modelling (Cost-Benefit Analysis, Meta-Analysis, Systematic Reviews);
  • Conducted several graduate and contract statistical analysis projects from CRO/biostatistics fields;
  • Producing Statistical Analysis Plans (SAPs) and liaising with support Statistical Programmers to review related Tables, Figures and Lists (TFLs);

Typical sections covered in the SAPS are description of study (primary and secondary outcomes); protocol deviations and changes in SAP; statistical considerations, study subjects and datasets analysed, including demographic and baseline characteristics; study effects (safety, efficacy, minor and adverse side-effects); and early study termination

  • Contributing to protocol development and reviewing Clinical Report Forms (CRFs);
  • Developing suitable randomisation codes for specific studies;
  • Performing relevant survival analysis, factor analysis, and Bayesian analysis for target clinical outcomes (using SAS Proc PH REG and Proc LIFETEST);
  • Performing mixed-effects and Generalized Linear Models (GLMs) using appropriate SAS procedures e.g. PROC MIXED and Proc GLM;
  • Leveraging deep knowledge of experimental design and biostatistics in applying suitable clinical statistical research techniques in specific disease areas (e.g. oncology, diabetes, depression, and dermatology), for key areas of Phase I-Phase III studies (drug efficacy and safety);
  • Understanding the regulatory regimes for clinical modelling of different therapeutic areas, for example oncology.

Statistical modelling and health economics

  • Designing economic models for early stage and final submission stages – techniques involved include meta-analysis of relevant research on statistical modelling in health economics, systematic reviews, applications to specific contract research contexts
  • Using an innovative Hybrid Statistical and Problem-Solving Model © to produce innovative data-driven evaluations of the cost-benefit trade-offs among competing models;
  • Expanding clients’ knowledge of health technology assessments (HTA) considering detailed requirements engineering and business analysis (case-based, future state considerations, for example);
  • Interpreting and training data modelling teams on the economic models;
  • Combining insights from other disciplines with economic models to optimize requisite decision-making for specific client contexts, namely Operational Research (OR), theory of constraints, scenario modelling, creative problem solving, and marketing research.

Confidential research on KEYTRUDA German HTA Submission

  • Assisted with the statistical aspects of the KEYTRUDA studies; KEYTRUDA (also known as pembrolizumab or MK-3475);
  • It involves large deliverables (over 1150 outputs) to address the benefit assessment of the drug per current approval status (therapeutic indication, dosage, contra-indications, concomitant treatment, for example), and on patient-relevant endpoints, as defined by the Social Codebook (mortality, morbidity, quality of life, adverse events);
  • For mortality and morbidity, reports were generated on time points (months), subgroups and sensitivity analysis;
  • For safety, the endpoints considered ae: time points (months), subgroups, exact p-values, RR, time-to-event;
  • For Patient Reported Outcomes (PRO), the endpoints considered are: time points (months), subgroups, MNAR and MAR
  • Kaplan-Meier curves were also used.
  • Contributed to the development of the SAP;
  • Worked on the health economic model by: a) generating patient data for the indicated label population to validate the model; b) analysing survival data for different endpoints to provide parametric risk functions to estimate risk of events in the health economic model for CAD/PAD population; and
  • Project managing part of the deliverables such as HTA metrics summary, health economic model validation and some endpoints

Confidential contract research on modelling pension data (2014)

  • A client company purchased an expensive set of annuities data from the Australian market, containing immediate and deferred pensions on different pension schemes. Based on confidential contract research arrangements, I did the following statistical modelling for them;
  • Customer actuarial and behavioral modelling involving patterns, life-style characteristics, and preferred insurance options on different pension schemes;
  • Developed a pricing strategy considering retirements, incentives, early retirement factors at different censoring ages (60, 65, for example), penalties and percentage of full benefits allowed, salaries, etc.;
  • Used knowledge of actuarial lifetables (Bayes’, actual and expected), GLM, survival analysis, data cleansing and smoothing, to perform business-critical predictive modelling of drivers of actual and predicted retirements;
  • Performed suitable exploratory data analysis on the data behavior and correlations among the study variables, and additional inferential statistics to gauge the significance of contributing variables to the best models: correlation matrices, one-way ANOVA, multivariate statistical analysis;
  • Leveraged my knowledge of the Actuarial Control Cycle in linking the modelling results to requisite industrial experience – premium pricing and pricing of the pension portfolios, for example;
  • Liaised with staff with deeper industrial experience to adjust the results to prevailing market expectations;
  • Used the SAS ODS (Output Delivery System) to generate requisite reports (PROC TABULATE, etc.);
  • The modelling was relatively full-spectrum analytics, which pooled vital elements of Patrick‘s training in Statistics, Business and Finance, with deep-seating critical literature review and meta-analysis of cutting-edge research in the project themes.
  • A conservative estimate of the life-time competitive advantage of the modeling is circa (£20m over a 5-year period, which more than compensates for the cost data and the contract research);
  • Finally, made strategic presentations to the project sponsors within the company, and trained their line staff on the results on hand-over at the end of the engagement.

Enquiries

For more details, please contact us at:

Email:  info@oseluxworldhero3e.com or through Tel. no:  +447772632150 to speak to the Director of Research and Enterprise Development, Dr Patrick Oseloka EZEPUE