Hello, I'm Lindiwe

Data Scientist | Developer | Insight Creator

View My Work Get In Touch

About Me

I thrive on discovering unconventional solutions to complex challenges. By weaving through the rich tapestry of data, I uncover patterns that transform obstacles into opportunities. My career path winds through diverse landscapes—each turn connected by a singular thread: my ability to craft solutions that resonate. What sets me apart? A storyteller's touch. I don't just present data; I guide audiences through narratives that captivate and conclude with clarity, bringing diverse minds together toward a common vision.

My journey in tech has been driven by curiosity and a desire to solve complex problems through elegant solutions.

Lindiwe Profile Picture

Featured Projects

SA Crime

SA Crime Intelligence Report

A rigorous data science interrogation of South Africa's Q3 2025/26 crime statistics, cross-referenced with two decades of SAMRC femicide surveillance data. Applies correlation analysis, anomaly detection, OLS regression, and ARIMA forecasting to answer whether the reported 3.2% rape decrease reflects genuine progress — or fewer women trusting a broken system.

Python ARIMA Statsmodels Scikit-learn SAPS Data SAMRC
Ore Mining

Mining Quality Intelligence

A full-stack ML system predicting silica concentrate quality failures in an iron ore flotation plant hours before they occur. Trained on 737,453 hourly sensor readings using XGBoost and a Neural Network with SMOTE balancing, business-optimal threshold tuning, and a live Streamlit dashboard with real-time GREEN / AMBER / RED operator alerts.

Python XGBoost TensorFlow Streamlit SMOTE Plotly
Tyla Grammy

Tyla Grammy Sentiment Analysis

A powerhouse NLP study dissecting the global phenomenon of Tyla's historic Grammy triumph, revealing a staggering 82% positive sentiment across 214 authentic fan reactions spanning 20+ countries. This isn't just data—it's a compelling narrative of cultural impact, leveraging cutting-edge dual-model sentiment validation (VADER + TextBlob) to separate genuine enthusiasm from algorithmic noise. Stunning interactive visualizations transform raw YouTube commentary into actionable brand intelligence, pinpointing high-growth markets and quantifying the ROI of award season visibility. The methodology serves as a scalable blueprint for music industry stakeholders seeking data-driven competitive advantages in an oversaturated market. Ultimately, this project showcases how sophisticated technical proficiency translates into deliciously strategic business insights for the global creative economy.

Python NLP Sentiment Analysis YouTube Data API v3 Regex Pandas VADER TextBlob Seaborn Matplotlib WordCloud
Diamonds Are Forever

Diamonds Are Forever

Full-stack ML study investigating the structural disruption of the global diamond industry — quantifying the 73.8% lab-grown price collapse since 2020, identifying five distinct buyer archetypes via K-Means clustering, and forecasting retail diamond prices with R²=0.9997 (Random Forest, RMSE=$136) — exceeding the published academic benchmark. Deployed as a live 4-page Streamlit intelligence dashboard.

Python Random Forest XGBoost K-Means SMOTE PCA Streamlit Plotly

Skills & Technologies

Data Analysis

Python
SQL
Excel

Machine Learning & AI

Machine Learning
Deep Learning
NLP

Data Visualization & BI

Power BI
Tableau
Matplotlib/Seaborn

Get In Touch

Email

sl.songelwa@hotmail.co.za

Location

Gauteng, South Africa