Skip to content
View reevesgonah's full-sized avatar

Block or report reevesgonah

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
reevesgonah/README.md

Profile views


🧭 About Me

I'm a graduate mechanical engineer based in Nairobi, Kenya, adding data science and ML to my toolkit to build real-world solutions where engineering and AI meet. Trained through the Data Science Bootcamp and Data Analytics course at Moringa School, and currently getting hands-on experience with real organizational data through data analysis, reporting and dashboarding work at Total Care Academy.

  • πŸ‡°πŸ‡ͺ Just shipped the Kenya Conflict Intelligence System - a political violence severity prediction capstone project
  • βš™οΈ Background in mechanical engineering - I bring domain knowledge into projects like predictive maintenance modelling
  • πŸŽ“ Looking to bridge the gap between Machine Learning and AI systems and practical large scale usability in industry.
  • πŸ“Š Data analysis, reporting and dashboarding work at TCA Nairobi, Kenya.
  • 🌱 Also into tennis and creative hands on projects like art, music and fashion design.

πŸ› οΈ Skills

Languages & Core Libraries

Analysis & Explainability

BI & Visualisation

Tools


πŸš€ Featured Projects

πŸ‡°πŸ‡ͺ Kenya Conflict Intelligence System

A two-stage hurdle model predicting the severity of political violence across Kenya's 47 counties at weekly resolution, trained on 16,600+ ACLED county-week records enriched with WorldPop density and election-cycle features.

  • Random Forest selected over XGBoost/TabNet for best precision-recall balance (Stage 1 F1 64.3%, Recall 86.4%)
  • Live Streamlit dashboard with SHAP explainability and a county-level vulnerability map
  • Automated weekly ACLED refresh via GitHub Actions

Python scikit-learn XGBoost TabNet SHAP Streamlit πŸ”— Live App

Predictive Maintenance β€” CNC Machine Failure Detection)

A binary classifier predicting CNC machine failure from real-time sensor data (temperature, torque, tool wear, rotational speed), shifting maintenance from reactive/time-based to data-driven scheduling.

  • Domain-informed feature engineering (Mechanical Power, Temp Difference) drawing on ME background
  • SMOTE-balanced training to handle a 3.4% failure rate
  • Optimised Decision Tree: 77.8% recall, 96.8% accuracy

Python scikit-learn imbalanced-learn

Automated Sentiment Analysis for Tech Brand Monitoring

An NLP pipeline classifying ~9,000 tweets about Apple and Google products as Positive, Negative, or Neutral, built for a fictional social-listening client whose manual labelling no longer scaled.

  • Five models compared iteratively; tuned binary LinearSVC hit macro F1 = 0.7694
  • SHAP + LIME explainability confirmed the model learns real lexical patterns
  • Unsupervised KMeans topic clustering on TF-IDF/SVD embeddings

Python scikit-learn NLTK Keras SHAP LIME

More on the way

Additional Power BI, Tableau, and Excel dashboard projects (sales analytics, aviation risk, box-office ROI) are documented in my full portfolio - plus new work coming out.


πŸ’Ό Experience

Data Analyst (Part-Time) β€” Total Care Academy, Pangani β€’ Consolidating student, financial, and academic records from multiple sources into structured datasets for ongoing analysis and reporting. β€’ Designing and building Power BI dashboards tracking financial trends, student performance, and teacher metrics; producing termly reports presented to relevant staff and directors board.


πŸ“ˆ GitHub Stats


πŸ“« Reach me: open to collaborations on data science and ML projects.

Pinned Loading

  1. Automated-Sentiment-Analysis-for-Tech-Brand-Monitoring Automated-Sentiment-Analysis-for-Tech-Brand-Monitoring Public

    End of phase (4) applying natural language processing (NLP) techniques to classify sentiment in tweets discussing technology brands and products.

    Jupyter Notebook

  2. Aviation-Safety-Risk-Analysis-for-Aircraft-Procurement Aviation-Safety-Risk-Analysis-for-Aircraft-Procurement Public

    End of phase project involving python fundamentals, pandas, data cleaning, manipulation and visualization.

    Jupyter Notebook

  3. Data-Driven-Insights-into-movie-box-office-success- Data-Driven-Insights-into-movie-box-office-success- Public

    Group project for moringa dsf-pt14, collaborative phase 2 project. Providing a new movie studio insights to allow better decision making. Eliud Kibet was also a contributor but unable to push, chan…

    Jupyter Notebook

  4. kenya-political-violence-risk-prediction kenya-political-violence-risk-prediction Public

    Forked from KimutaiHub/kenya-political-violence-risk-prediction

    Kenya has experienced recurring political violence since 1991. Rich ACLED conflict data and Kenya Census Population & Density Data exists, yet little structured ML work has been applied to predict …

    Jupyter Notebook

  5. Optimizing-Industrial-Operations-Through-Predictive-Maintenance-Modeling Optimizing-Industrial-Operations-Through-Predictive-Maintenance-Modeling Public

    End of Phase (3) project implementing the use of machine learning techniques (Logistic Regression and Decision Tree Classifiers) for prediction and solving a business problem.

    Jupyter Notebook

  6. KimutaiHub/kenya-political-violence-risk-prediction KimutaiHub/kenya-political-violence-risk-prediction Public

    Kenya has experienced recurring political violence since 1991. Rich ACLED conflict data and Kenya Census Population & Density Data exists β€” yet little structured ML work has been applied to predict…

    Jupyter Notebook 2 5