Smaila Amoanu

Statistician specializing in time series, calibration, and interpretable ML. I build models that are not only accurate but diagnosable and deployable—linking theory to production-quality decisions in health and risk.

Interest Areas

• Information Technology (I.T)
• Education & Research
• Writing and Blogs
• Entrepreneurship

Research Interest

• Probability & Measure

• Statistical Inference (GLR/likelihood)

• Time Series (ARMA/GARCH, VAR)

• Biostatistics & Epidemiology
• Financial Mathematics

• Computational Statistics

Welcome to my YouTube journey! Back in 2018, I embarked on an extraordinary adventure on YouTube, driven by my passion for sharing knowledge about Database Management with the world. It's been an incredible ride, filled with moments of genuine connection and joy, especially when I receive heartfelt testimonials from viewers whose lives have been positively impacted by my tutorials.

My channel is a dedicated space where I freely share tutorials on essential topics such as Microsoft Access databases, Microsoft Excel, and SQL. I firmly believe that empowering individuals with the skills to understand, store, manage, and analyze data is pivotal for informed decision-making and making a real difference in the world.

If you're eager to dive into the world of data management and unlock its potential, I invite you to explore my playlists by clicking the Tube placeholder.

EXPLORE MY SAMPLE PROJECTS

This system is a game-changer, offering robust functionality to streamline various aspects of school administration. From managing student records, handling school fees, facilitating examinations, and maintaining staff information to tracking attendance, processing payroll, and generating efficient reports with insightful visualizations – we've got it covered.

Curious to explore more about this transformative solution? Click on the placeholder to experience the full demo firsthand.

Designed as a multi-user database, this system is ideal for churches of all sizes, offering robust features to support various aspects of church management. Whether it's tracking member information, managing contributions, organizing events, or overseeing volunteer coordination, this system has you covered.

Built to be shared over a LAN or Extranet, it ensures seamless collaboration among church staff and volunteers.

The sysem is designed to simplify inventory management, this system empowers you to efficiently store and track products across multiple locations within your warehouse.

Versatile in its application, this inventory system isn't limited to warehouses alone. It's equally effective for businesses of all kinds that deal with inventory management.

Whether you're overseeing a bustling warehouse operation or managing inventory for your business, this system provides the tools you need for seamless tracking and organization.

The system is designed with simplicity in mind, our POS software is a versatile solution suitable for a wide range of businesses, including shops, supermarkets, pharmacies, and both retail and wholesale operations.

With its intuitive interface and robust functionality, our POS system streamlines transactions and enhances efficiency in your business operations. Whether you're processing sales, managing inventory, or tracking customer data, our POS software is equipped to handle it all.

For information on the latest versions, do not hesistate to reach out.

In this project, I utilized the freMPL3 dataset to train predictive models for insurance technical price estimation. I employed KNN, Random Forest, and XG Boost algorithms for regression tasks, evaluating performance with RMSE and R-squared metrics. Preprocessing techniques like scaling and target encoding were applied, and model interpretation was facilitated through Partial Dependence Plot (PDP) and SHAP. Benchmarking revealed Random Forest and KNN as superior in explaining variability and predicting technical price, with contributions varying between genders. This comprehensive approach ensures accurate and interpretable insurance pricing decisions.

Controlling malaria in Geita is challenging due to widespread antimalarial drug resistance, exacerbated by economic constraints limiting the efficient deployment of interventions like case management, bednets, indoor residual spraying, and vaccines. The study aims to optimize intervention strategies by identifying the most cost-effective combination through OpenMalaria simulations, utilizing historical data from the region to evaluate effectiveness compared to VecNet baseline models.

Access report here

I quantified how much IRS contributed to reducing malaria transmission across Ghanaian districts by building complementary deterministic and stochastic transmission models. Parameterized with Malaria Atlas Project geospatial prevalence inputs, the models compared counterfactual scenarios (with/without IRS) and ran sensitivity analyses on key entomological and operational parameters. The results informed district-level targeting and cost-effectiveness guidance for malaria control.

I modeled the short- and long-run relationships between Ghana’s government expenditure and GDP using Vector Autoregression (VAR) and Vector Error-Correction Models (VECM). The workflow included stationarity testing (ADF/KPSS), lag selection (AIC/BIC), Johansen cointegration, estimation of VAR/VECM, and policy-oriented interpretation via impulse response functions (IRFs) and forecast error variance decomposition (FEVD)—clarifying how spending shocks propagate to output over time.