Data Analytics and Operations Research

Harnessing the power of data analytics and operations research to optimize operations within various sectors including finance, retail, and healthcare.

New Product Order Quantity Machine Learning Model

Part of a consulting project with Abyat Megastore

  • The research addresses the challenge of forecasting sales for products yet to be launched, which inherently lack historical sales data.

  • Traditional time-series forecasting techniques are inapplicable due to the absence of historical data.

  • The study aims to harness machine-learning algorithms, specifically drawing from sales data of comparable products launched previously.

  • A key consideration is the influence of external factors like seasonality and promotions on new product sales.

  • The research methodology is inspired by notable works such as Smirnov and Sudakov (2021).

  • The study underscores the complexities introduced by seasonality and promotions, emphasizing the need for a nuanced model.

  • The research leverages advanced data analytics, particularly machine-learning, to enhance forecast accuracy for products without prior market presence.

Optimizing Patient Flow at Kuwait Cancer Center Outpatient Department

As the principal investigator of a Kuwait Foundation for Advancement of Science grant

  • The project centers on the critical task of optimizing patient flow within the outpatient department of Kuwait Cancer Center.

  • A rigorous data analysis was undertaken to accurately parameterize the model, ensuring its relevance and applicability.

  • Comprehensive interviews with clinical staff and doctors were conducted, providing invaluable insights into the existing processes and potential areas of improvement.

  • The study employed advanced process mining procedures, shedding light on the intricate patient flow dynamics and identifying bottlenecks.

  • As a direct outcome of the research, significant enhancements were made to patient scheduling protocols, ensuring a smoother and more efficient patient experience.

  • Clinic schedules were revamped based on the findings, leading to optimized resource allocation and reduced patient wait times.

  • The study's results also served as a catalyst for service expansion, highlighting the need and providing the impetus for broadening the center's service offerings.

  • Overall, the project stands as a testament to the transformative power of data-driven decision-making in healthcare settings, leading to tangible improvements in patient care and operational efficiency.

The Weekend Effect on Rewards-Based Crowdfunding

With Humoud AlSabah

  • The projects's primary objective is to discern the influence of the day of the week on the success rates of crowdfunding campaigns.

  • Utilizing a comprehensive dataset spanning from 2009 to 2018, the analysis encompasses over 185,000 crowdfunding projects.

  • A significant finding is the diminished likelihood of success for campaigns initiated during weekends compared to those launched on weekdays.

  • Contrarily, the day of initiation does not impact the success rates of projects launched on national holidays, distinguishing the "Weekend Effect" from a broader holiday effect.

  • Practical implications suggest that if campaigns are initiated on weekends, aligning their deadlines to weekends might be advantageous.

  • The observed patterns remain consistent across various mixed-effects statistical models, ensuring the robustness of the findings.

  • The study also accounts for potential confounding variables, such as the project's initiation year, its featured status on the crowdfunding platform, and the set funding goal, ensuring a comprehensive analysis.

Moral Hazard in Online Peer-to-Peer Lending

With Humoud AlSabah

Published in Applied Economics (link)

  • The project asks if changes in online peer-to-peer lending could make it riskier for lenders due to potential dishonest actions by the lending platform.

  • To answer this, we created a model that looks at how online lending platforms decide who gets a loan and at what interest rate.

  • They discovered that when there aren't many trustworthy borrowers, these platforms might be less strict in checking who they lend to. This can lead to loans being given to less reliable borrowers, which isn't good for those lending money.

  • This suggests that the rules currently in place might not be doing enough to protect people who lend money on these platforms.

  • One solution could be for regulators to make sure these platforms have a personal stake in the loans, ensuring they're more careful about who they lend to.

Management Dashboards in Primary Care: Supporting Data-Driven Management

With Dr. Huda Aluwaisan, Eng. Hazza Alotaibi

Presented at the World Organization of Family Doctors (WONCA) Conference 2022

  • Traditional performance reporting in healthcare is often static and lacks transparency.

  • Dashboards offer a transformative approach, converting complex data into interpretable visuals and streamlining decision-making.

  • The dashboard we developed in cooperation with Yarmouk Primary Care Center exemplifies a pioneering shift in healthcare management.

  • The workshop delineated the intricate process of dashboard design, underscoring the synergy between clinical managers and data engineers.

  • Yarmouk's management found the dashboard pivotal in identifying operational anomalies and their root causes.

  • The dashboard's granularity allows insights at multiple operational tiers, from patient-centric to specialty-clinic perspectives.

  • Yarmouk's innovative approach underscores the potential of dashboards in revolutionizing clinical management.

  • The workshop aimed to disseminate Yarmouk's groundbreaking insights, emphasizing the imperative of data-driven decision-making in modern healthcare.

Previous

Global Health & Payment Reforms in Low- and Middle-Income Countries