34 Projects. All Live. All Documented.

This page highlights representative live work from our wider portfolio of 34 deployed projects across 7 industries. Additional live products and simulator listings are linked from the Products page. Everything shown here is documented honestly โ€” including the failures, the data leakage fix, and the models that underperformed.

๐Ÿ“Š ML Models ๐Ÿ› ๏ธ Simulators ๐Ÿ“š LMS ๐Ÿซ EdTech Tools
34
Total Live Projects
7
Industries Served
7
ML Models Deployed
11
Data Tool Simulators
โ‚ฆ0
CBT Pro Budget
๐Ÿ“Š
Machine Learning & Data Science Models
End-to-end pipelines ยท All live on Streamlit ยท @cssadewale / @hmgtechnologies
7 Models
Live ยท NLPText Classification ยท 3MTT Capstone
TruthLens โ€” Fake News Detector
ProblemNot every student or reader can fact-check at speed in today's information environment.
SolutionFull NLP pipeline: cleaning โ†’ tokenisation โ†’ lemmatisation โ†’ TF-IDF (20K features) + VADER sentiment โ†’ XGBoost. Formally assessed 3MTT Capstone 2.
AUC 0.9393Acc 86.75%12,999 articles
XGBoostTF-IDFNLTKStreamlit
LiveRegression ยท Workplace Wellbeing
NeuroWell โ€” Employee Burnout Rate Predictor
ProblemPredict employee burnout before productivity and health collapse โ€” an early warning, not a lagging consequence.
SolutionGradient Boosting ยท 108 GridSearchCV configs ยท Mental fatigue (r=0.878) = 85.9% of model importance ยท joblib serialisation.
Rยฒ 0.855RMSE 0.07222,750 records
Gradient BoostingGridSearchCVjoblibStreamlit
LiveBinary Classification ยท Banking
Bank Customer Churn Prediction
ProblemIdentify customers about to leave before they do โ€” turning reactive retention into proactive strategy.
SolutionGradient Boosting on 10,000 bank records. Multiple classifiers compared on precision-recall tradeoffs. Quantified retention recommendations.
F1 0.609AUC 0.86810,000 records
Gradient BoostingScikit-learnStreamlit
LiveBinary Classification ยท Insurance
Insurance Claim Prediction
ProblemPredict fraudulent or high-cost insurance claims before processing to stop revenue loss.
SolutionRandom Forest + SMOTE + GridSearchCV + SHAP + FastAPI on 7,014 records. CV F1 ~0.79. 6 business recommendations. SHAP explains every prediction.
CV F1 ~0.797,014 recordsSHAP + FastAPI
Random ForestSHAPFastAPIStreamlit
LiveBinary Classification ยท HR Analytics
Algorithmic Staff Promotion โ€” Yakub Trading Group
ProblemRemove bias from HR promotion reviews. Flag high-potential employees with data, not gut feel.
SolutionRandom Forest + GridSearchCV + SHAP on 38,312 records. Fairness analysis via SHAP. Objective, data-backed promotion recommendations.
AUC 0.89138,312 recordsSHAP Fairness
Random ForestSHAPGridSearchCVStreamlit
LiveBinary Classification ยท Financial Inclusion
Income Level Prediction
ProblemPredict income bracket to power financial inclusion and credit scoring for underbanked populations.
SolutionUCI Census dataset (48,842 records). 5-model comparison with SMOTE. Full pipeline: encoding, scaling, imbalance treatment.
48,842 records5-Model ComparisonSMOTE
Random ForestSMOTEStreamlit
LiveMulti-class Classification ยท Logistics
๐Ÿšš SwiftChain Delivery Delay Prediction
ProblemClassify 15,549 logistics records across three delivery outcomes before dispatch.
NotableTransparently documents a data leakage fix in the notebook โ€” because honest engineering matters more than clean-looking numbers.
Acc 62%F1 0.57915,549 orders
Gradient BoostingSMOTEStreamlit
๐Ÿ“š
HMG Academy LMS โ€” Learning Platforms
Data science curriculum + Nigerian secondary school curriculum ยท @hmgacademyhub
11 Platforms
๐Ÿ—„๏ธ
SQL Learn

Interactive SQL learning โ€” from fundamentals to advanced queries. Built for data analysts and scientists.

โ†— Open Platform
๐Ÿš€
Streamlit Learn

Learn to build and deploy live data apps with Streamlit โ€” from notebook to web app.

โ†— Open Platform
๐Ÿ“Š
Excel Learn

Comprehensive Excel โ€” from basic spreadsheets to advanced formulas, pivot tables, and dashboards.

โ†— Open Platform
๐Ÿงฌ
Data Science Learn

End-to-end data science curriculum โ€” EDA, feature engineering, model building, and deployment.

โ†— Open Platform
๐Ÿ“ˆ
Data Analysis Learn

Practical data analysis from raw data to insight โ€” Excel, SQL, Python, and visualisation.

โ†— Open Platform
๐Ÿ
PyDataFlow

Python for Data Science โ€” structured from first print() to deployed ML model for absolute beginners.

โ†— Open Platform
๐ŸŽจ
VizLearn

Data Visualisation in Python โ€” Matplotlib, Seaborn, and Plotly for clear, communicative charts.

โ†— Open Platform
๐Ÿค–
ML Lab

Machine Learning curriculum โ€” classification, regression, NLP, and model evaluation with real datasets.

โ†— Open Platform
๐ŸŽ’
Junior Class

Nigerian JSS1โ€“JSS3 curriculum โ€” all core subjects for junior secondary students.

โ†— Open Platform
๐ŸŽ“
Senior Class

Nigerian SSS1โ€“SSS3 curriculum โ€” Science, Arts, and Commercial streams. WAEC/NECO focused.

โ†— Open Platform
๐Ÿงญ
QueryPilot

Intelligent SQL query assistant โ€” helping learners write, debug, and understand SQL with guided support.

โ†— Open Platform

"Every project on this page started not with a framework or a dataset โ€” but with a problem I watched real students, teachers, businesses, or organisations struggle with. The classroom, the boardroom, the community โ€” these are the briefs. We don't wait for the perfect environment. We use what we understand clearly, leverage AI confidently, and build the solution the problem actually needs."

โ€” Adewale Samson Adeagbo ยท Founder, HMG Technologies ยท AI-Augmented Solutions Developer
Work With Us See Our Services