Projects
Below are selected research projects supported by multiple grants reflecting interdisciplinary work in Data Science and Artificial Intelligence.
📁 Please finds full list of publications and patents.
Empowering Cloud Computing for Non-image-based Diabetic Retinopathy Screening by Designing an EHR-oriented Incremental Learning Framework
Summary: This NIH-funded project aimed to address the scalability and adaptability challenges of AI models for diabetic retinopathy (DR) screening in primary care. By leveraging cloud computing, the project developed an EHR-compatible incremental learning framework that could continuously learn from new patient data without retraining from scratch. This non-image-based approach utilizes routine lab results and comorbidity data—making it more accessible, especially in rural or resource-limited settings—while achieving comparable accuracy to image-based DR screening tools.
(2023) NIH-NEI: NOT-OD-23-070 — Role: Researcher
Harnessing Tensor Information to Improve EHR Data Quality for Accurate Data-driven Screening of Diabetic Retinopathy with Routine Lab Results
Summary: Funded under an NIH R01 grant, this project explored tensor-based representations to model multi-dimensional EHR data more effectively. It focused on enhancing data quality and reducing noise and missingness, key barriers in clinical ML adoption. By applying tensor decomposition and imputation techniques, the research improved the reliability of DR prediction using only non-ophthalmic lab values, ultimately contributing to a more robust and interpretable screening tool for early-stage detection in routine clinical practice.
(2022) NIH-NEI: 5R01EY033861 — Role: Researcher
Application of Supervised Learning Algorithms in the Service Industry
Summary: This applied research project investigated the use of supervised machine learning—such as decision trees, support vector machines, and random forests—to enhance key performance indicators in Indonesia’s service industry. It focused on demand prediction, customer segmentation, and service optimization, providing data-driven recommendations that improved operational workflows and customer satisfaction across hospitality, retail, and health sectors. The project also emphasized model interpretability and deployment feasibility for non-technical stakeholders.
(2022) PUTI No. NKB-1337/UN2.RST/HKP.05.00/2022 — Role: Co-PI
Application of Big Data Analytics to Improve the Quality of Risk Analysis
Summary: This project sought to improve enterprise-level risk management using big data techniques. It collected and analyzed large-scale, heterogeneous datasets—ranging from financial transactions to operational logs—to identify latent risk factors and predict organizational vulnerabilities. Techniques such as anomaly detection, clustering, and risk scoring were used to create early warning systems, helping decision-makers proactively manage operational and financial risk.
(2020) PUTI No. NKB-1702/UN2.RST/HKP.05.00/2020 — Role: Co-PI
Development of Time Series Forecasting Models Using a Data-Driven Approach
Summary: This research focused on designing and validating time series forecasting models tailored to dynamic business environments. The team applied ARIMA, SARIMA, and LSTM-based deep learning approaches to predict trends in sales, energy usage, and other KPIs. Emphasis was placed on model interpretability, seasonality detection, and real-time prediction capability, enabling data-driven decision support for industries undergoing digital transformation.
(2020) PUTI No. NKB-1133/UN2.RST/HKP.05.00/2020 — Role: Co-PI
Escalating Company Performance Using a Data Mining Approach
Summary: By analyzing large internal datasets, this project identified performance bottlenecks and improvement opportunities within Indonesian enterprises. Using association rule mining, clustering, and classification algorithms, the research provided actionable insights for enhancing productivity, workforce allocation, and strategic planning. The approach empowered companies to shift from intuition-based to evidence-based decision-making.
(2020) PUTI No. NKB-1070/UN2.RST/HKP.05.00/2020 — Role: Co-PI
Application of Data Mining to Accelerate Performance Improvement in the Manufacturing and Service Sectors
Summary: This interdisciplinary project bridged data science with industrial engineering to optimize workflows in both manufacturing and service sectors. Techniques such as k-means clustering, decision trees, and regression analysis were applied to real-world operational data, improving throughput, reducing waste, and enhancing customer experience. The project emphasized customized solutions aligned with sector-specific constraints and challenges.
(2019) PIT-9 No. NKB-0061/UN2.R3.1/HKP.05.00/2019 — Role: Co-PI
Improving Operational Effectiveness and Efficiency Through Data-Driven Method Approaches
Summary: This initiative aimed to build internal analytics capability within organizations through a structured data-driven methodology. It involved the integration of KPIs, continuous monitoring dashboards, and predictive models to optimize resource usage, cut delays, and enhance service delivery. Several pilot case studies demonstrated tangible improvements in turnaround times, quality assurance, and customer feedback metrics.
(2018) PITTA No. 2455/UN2.R3.1/HKP.05.00/2018 — Role: Co-PI
Enhancing the Competitiveness of Indonesia's Service Industry Through Big Data Analysis
Summary: This project focused on the strategic use of big data to foster innovation and competitiveness in Indonesia’s fast-growing service sector. It involved building data pipelines, performing sentiment and market trend analysis using social media and transactional data, and developing dashboards for performance tracking. Policy recommendations were made for government and private sectors to leverage data assets more effectively.
(2018) PITTA No. 787/UN2.R3.1/HKP.05.00/2017 — Role: Researcher
Application of Data-Driven Methods to Improve the Performance of State-Owned Enterprises
Summary: Targeting inefficiencies in Indonesian SOEs, this study applied statistical and machine learning models to benchmark performance, identify inefficiencies, and recommend restructuring strategies. It focused on improving accountability, profitability, and transparency. By integrating data from HR, finance, and operations, the project empowered state enterprises to adopt a performance-driven culture.
(2017) PITTA No. 788/UN2.R3.1/HKP.05.00/2017 — Role: Researcher
Achieving Electrical Energy Resilience in Indonesia Through Improving Reliability, Forecasting Quality, and Electricity Distribution Using Data Mining and Graph Theory
Summary: This early project combined graph theory and predictive analytics to improve the reliability of Indonesia’s electrical grid. The team developed models to predict outages, optimize load distribution, and enhance network robustness. Using network topology analysis and failure propagation modeling, the project provided new tools to support energy resilience planning in urban and rural areas.
(2016) PITTA No. 2132/UN2.R12/HKP.05.00/2016 — Role: Researcher
Economic Indicator Data Analysis Using Data Mining to Predict Company Performance and Financial Fraud in Order to Improve Competitiveness and the Investment Climate in Indonesia
Summary: This project utilized national economic indicators, such as inflation rates, interest trends, and company financial statements, to develop predictive models for corporate performance and fraud detection. Using decision trees and logistic regression, the study identified red flags for investors and policymakers, ultimately contributing to a healthier investment climate and enhanced corporate governance.
(2016) PITTA No. 2134/UN2.R12/HKP.05.00/2016 — Role: Researcher