Titulo Estágio
Implementation of Functional Data Analysis Tools in a Data Analytics Platform for Root Cause Analysis in Biopharmaceutical Manufacturing
Local do Estágio
Roche Diagnostics GmbH, Penzberg, Germany
Enquadramento
Root cause analysis plays a crucial role in identifying and resolving issues in biopharmaceutical manufacturing processes. However, current methods used for root cause analysis are often less effective in retrieving the relevant information contained in manufacturing process data. In this context, functional data analysis is a promising statistical methodology to overcome these limitations, as many biopharmaceutical manufacturing process characteristics are of functional nature (time series / longitudinal data).
One of the company’s goals is to reduce time and cost with most efficient root cause analyses. Among others, this can be achieved by standardising tools across global networks, and providing user-friendly access to data analysis tools. This Master’s thesis will integrate a functional data analysis (FDA) framework into an existing company internal data analytics platform and will apply FDA to a current business use case from biopharmaceutical manufacturing.
Objetivo
The main objective of this work is the development of an R-based framework that incorporates functional data analysis techniques. The framework will integrate with an existing in-house built data analytics platform, providing a standardised and user-friendly tool for root cause analysis. The workflow within the framework will encompass data pre-processing, functional data representation, functional regression modelling, and visualisation techniques.
The implementation of the R-based functional data analysis tools into the platform's architecture will be tested for compatibility, usability and accessibility for users within the global network of the company. The effectiveness and efficacy of the FDA framework will be evaluated with a real-world business use case involving time series data from various stages of a biopharmaceutical manufacturing process. What advantages does FDA offer in comparison with existing root cause analysis methods in terms of its ability to capture complex temporal patterns, uncover hidden relationships in the data, and ultimately, identify key influence factors contributing to process variations? What is the impact of the framework on reducing time and cost associated with root cause analysis?
Plano de Trabalhos - Semestre 1
1. Literature review on functional data analysis and its applications in biopharmaceutical manufacturing processes, in particular root cause analysis.
2. Familiarisation with theoretical aspects of functional data analysis, in particular functional principal component analysis.
3. Familiarisation with the inhouse-built data analytics platform and other computerised systems as well as the biopharmaceutical manufacturing processes.
4. Conduct interviews with stakeholders and users to develop an adequate solution.
5. Perform technical training to enhance proficiency in R programming language.
6. Start the workflow design: data preprocessing, functional data representation, functional regression modelling, and visualisation techniques.
7. Completing the Intermediate report.
Plano de Trabalhos - Semestre 2
1. Completing the workflow design: data preprocessing, functional data representation, functional regression modelling, and visualisation techniques.
2. Completing the development and implementation of an R-based framework for functional data analysis within the existing data analytics platform.
3. Application of the FDA framework to a business use case in biopharmaceutical manufacturing.
4. Evaluation of the FDA framework: comparison with other root cause analysis methods, time efficiency, cost reduction, standardisation, and user-friendliness.
5. Completing the Master's thesis.
Condições
Physical Workspace:
A dedicated workspace will be provided to carry out the research at Roche Diagnostics GmbH in Penzberg. This workspace will be equipped with a company’s computer and necessary software.
Collaborative Network:
Interdisciplinary collaboration with domain experts, stakeholders, and members of the research team will be encouraged.
Data Sources and Systems:
Access to and training for various computerised systems utilised within the biopharmaceutical manufacturing company will be provided.
Literature and documentation:
Access to training resources, literature and documentation will be available.
Observações
n/a
Orientador
Marco Simões
msimoes@dei.uc.pt 📩