Blog

Combining multiple ML models with computer vision techniques for traffic monitoring, extracting features using SIFT/SURF, and color histograms to achieve 85% accuracy in vehicle detection.

Implementing attention mechanisms with AFROXLMR for Luganda PII detection, achieving 96.3% accuracy. Using LIME and SHAP for model interpretation and bias detection.

Implementing and comparing SVM, KNN, and Naive Bayes models with sampling techniques, achieving 62.3% AUC score. Analyzing impact of undersampling vs oversampling on model performance.

Processing millions of reviews using MapReduce for sentiment scoring and product ranking. Implementing distributed computing with custom reducer functions for aggregation and analysis.

Analyzing disease patterns across bovine, caprine, and ovine populations using R's tidyverse and Highcharter. Creating interactive choropleth maps for geographical distribution visualization.

Creating temporal analysis of chat patterns and relationship dynamics using R Shiny. Implementing text mining for sentiment analysis and topic modeling of conversations.

Automating web scraping, cleaning and data consolidation for bond yield analysis, as opposed to manual download from Bank of Uganda website. Using R's rvest for extraction and dplyr for transformation.

Analyzing demographic correlations and trust factors in vaccine uptake using R's tidyverse. Creating interactive visualizations with Highcharter for stakeholder presentations.

Applying Feature Analysis and logistic regression to identify key factors influencing team gender composition. Quantifying impact of researcher role (38.1%), gender (23.5%), and budget (7.7%).

Measuring efficiency gains of two-chamber stoves showing 40% heat energy improvement. Analyzing adoption patterns and environmental impact using R for statistical analysis.

Reducing reporting time from 5 days to 30 minutes using Power BI and R Shiny. Implementing automated KPI calculations and real-time data updates for investment decision-making.

Implementing data lake architecture using R and SQL for research findings management. Optimizing query performance and creating automated data validation checks.