TOTobechi Obinwanne

Tobechi Obinwanne

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Tobechi Obinwanne has a total of 22 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    FAIRification on Trial? A Provenance and Reproducible Metadata Build Pipeline for Clinical Research

    by Tobechi Obinwanne, Johannes Darms, Juliane Fluck, Ulrich Sax

    Clinical trial registries often focus on the human user and do not focus on machine-actionable structures and means to detect and manage provenance, thus hindering large scale reliable reuse and verification. This study presents a FAIRification pipeline for EU Clinical Trials Registry metadata, which preserves provenance and enables the reproduction of results. When the pipeline is run on the trials uploaded between 2022 and 2025, it was able to convert 94.84% of trial metadata into NFDI4Health metadata model. The process demonstrates almost complete population of the target model’s mandatory properties. The process ensures that each transformation has W3C PROV-O lineage. As RO-Crates, outputs allow federated inclusion in catalogs and verifiably repeatable reuse. We demonstrate that trial metadata should be made available as reproducible digital objects rather than textual or permanent narratives. Same inputs should result in identical outputs, and transformations should be completely auditable.

  • Open Access

    Supporting Reanalysis and Reuse of Clinical Trial Data: A Case Study

    by Cora Burgwinkel, Han Chang Chiam, Ka Hin Tai, Jifan Wang, Mian Haider Ali, Salman Soleiman Fallah, Minoo Matbouriahi, Tobechi Obinwanne, Grigorius Papapostolou, Muhammad Riedha, et al.

    Background: Reproducing published findings from clinical trials is a critical component of scientific transparency, yet it remains a challenging and under-practiced task. Despite increasing emphasis on reproducibility and data reuse in research policies, few real-world examples exist where independent teams have reproduced complex analyses using clinical trial data. In this case study, the aim was to independently reproduce the key findings of a high-impact clinical trial on rectal cancer treatment using shared trial data. Method: We organized a multi-team datathon, where each team was provided with the same dataset and supporting material, and was tasked to reproduce the results of the CAO/ARO/AIO-04 trial, with optional additional analysis. We contacted the original investigators for data access and reuse, and consulted them to understand the study, clinically and scientifically. Results: Five teams used R or Python to reproduce the statistical results, and the corresponding scripts can be found on Gitlab. All teams reproduced the analyses for primary outcome disease-free survival (DFS). The key findings on DFS were consistently reproduced, reinforcing confidence in the trial main conclusions. Result robustness was investigated using a different analytical software or statistical models. Nevertheless, challenges were encountered when the supplementary materials were not easily identified. Minor reporting issues were noticed in the reproduced paper. Conclusion: Reproduction of a major oncology clinical trial confirmed the reliability of its main conclusions. Divergences highlighted reporting gaps, such as incomplete protocols and broken links, that future trials should address. This case study demonstrates the value of systematic reproducibility checks for clinical research transparency and challenges in data sharing for reproducibility.

  • Applications of Large Language Models (LLMs) in Business Analytics – Exemplary Use Cases in Data Preparation Tasks

    by Mehran Nasseri, Patrick Brandtner, Robert Zimmermann, Taha Falatouri, Farzaneh Darbanian, Tobechi Obinwanne


  • Open Access

    The Application of Machine Learning Algorithms in Predicting the Usage of IoT-based Cleaning Dispensers

    by Tobechi Obinwanne, Chibuzor Udokwu, Patrick Brandtner

    Internet of Things (IoT) based liquid cleaning dispensers are being increasingly used in public buildings for personal sanitation purposes. However, it is not always easy for facility managers to keep track of, as well as predict product usage. Most devices deployed at facilities still require the facility/building manager or staff at the facility, to check the devices from time to time. In recent years, the need to effectively utilize these devices as well as anticipate usage rates has become necessary because the time lag between refilling the dispensers and their being out of service can pose health risks. This paper thus explores how machine learning (ML) algorithms can be applied to improve the availability of IoT-based liquid cleaning dispensers. The goal of the paper is to apply machine learning in predicting the daily usage volumes of cleaning solutions, thereby increasing the efficiency of the cleaning dispensers. The paper compares different machine learning algorithms to determine the best algorithm for predicting the usage patterns of IoT-based cleaning dispensers, thereby, develops a predictive model that can be applied to improve the availability of cleaning products in IoT-based dispensers. The results of the analysis show that the Random Forest algorithm performed best among the evaluated models using regression performance measures. Hence, ML algorithms can be applied to help building or sanitation managers improve the availability of cleaning products in IoT-based cleaning dispensers, ultimately improving the user experience.

  • Open Access

    Data Preprocessing in Supply Chain Management Analytics - A Review of Methods, the Operations They Fulfill, and the Tasks They Accomplish.

    by Tobechi Obinwanne, Chibuzor Udokwu, Robert Zimmermann, Patrick Brandtner

    Data preprocessing is thought of as one of the most important steps in data analytics. This is especially true for the field of Supply Chain Management (SCM), in which the handling of huge data sets is the norm. Data preprocessing consists of multiple tasks, operations, and methods. Thus, this research focusses on identifying the specific data preprocessing tasks in SCM analytics, the operations used to solve them, and the methods used to meet the goals of the respective operations. To this end, a literature review, covering literature from 2011 to 2022, was conducted to analyze documented approaches to data preprocessing in SCM. The resulting overview presents the interrelationship between data preprocessing tasks, data preprocessing operations, and data preprocessing methods in SCM analytics. Results indicate that data transformation seems to be a commonly investigated task in SCM related data preprocessing, while data integration represents an area requiring further research. Furthermore, Principal Component Analysis (PCA), was found to be the most common method across the single tasks of data preprocessing, further highlighting the importance of transforming data by manipulating the features into a form such that when analytics algorithms are applied, they will give optimal results. This research hence presents researchers and practitioners a point of reference to identify the specific data preprocessing method used for specific data preprocessing operations in order to fulfill a specific data preprocessing task.