I’m Sheyin Avong, an MSc Applied Data Science student at UCLan with a BSc (Hons) in Computing (2022). I build end-to-end analytical solutions: explore and clean data, develop and evaluate models, and ship lightweight apps and automations so results are usable in the real world.

What I bring

  • Data analysis & wrangling: clear EDA, feature engineering, and reproducible notebooks in Python/R; confident with SQL for querying and joins.
  • Modeling: classical ML (e.g., logistic regression, random forest), computer vision deep learning (ResNet18, MobileNetV2, YOLO, VGG16) with robust evaluation (ROC-AUC, confusion matrices).
  • NLP: end-to-end text pipelines — cleaning, tokenization, vectorization (TF-IDF/embeddings), and transformer-based tasks like text classification, named-entity recognition, and summarisation using spaCy/Hugging Face.
  • Engineering mindset: production-ready scripts and small web tools/APIs with Flask; config/logging, batch jobs, and Git-based workflows.
  • Communication & visuals: clear write-ups, tidy plots/dashboards, and practical documentation so others can run and extend the work.

Selected work

Gesture-Controlled 2D Driving Game (YOLOv8)

Real-time hand gestures via webcam. Compared ResNet18+MediaPipe, MobileNetV2, and YOLOv8 (demo shows YOLOv8).

PythonOpenCVPyTorch YOLOv8Pygame
Watch demo Project page

Concrete Crack Classification & Segmentation

Traditional + deep learning: HOG+SVM, VGG16, U-Net, Otsu thresholding for crack detection and masks.

PythonComputer VisionHOG+SVM VGG16U-Net
View notebook Project section

Building Data Analysis

Energy/usage patterns with reproducible EDA and visualisation; insights you can act on.

PythonPandasVisualisation
View notebook Project section

Housing Affordability Trends

Data wrangling + regression with scikit-learn; clean plots and narrative.

Pythonscikit-learnEDA
View notebook Project section

NLP Mini-Projects

Prototype pipelines for NER, role classification, and summarisation with spaCy/Transformers.

spaCyTransformersTF-IDF/Embeddings
See examples

More details and downloadable notebooks live on the Projects page.

Freelance & applied work

  • Automated certificate generation system (Python/Flask): a web workflow that batch-creates training certificates (templating, Dropbox integration, logs, duplicate handling).
  • Business automation: customised chatbots, workflows, and sites in Go High Level for small businesses.
  • Teaching: introduced students to web development and Python fundamentals.

Education & recognition

  • MSc Applied Data Science (UCLan): Machine Learning, Visual Information Processing, Programming with Data, Internet of Things, Big Data Analytics & Visualisation.
  • BSc (Hons) Computing (UCLan): 2:1; School Award for Best Project.

Skills & tools

  • Languages: Python, R, SQL, HTML/CSS/JS, C#
  • Cloud Platforms: AWS & Microsoft Azure
  • Data/ML: Pandas, NumPy, scikit-learn, PyTorch, OpenCV,
  • NLP: spaCy, Hugging Face Transformers, NLTK, regex; TF-IDF/embeddings; classification/NER/summarisation
  • Web & tooling: Flask, Laravel; Git/GitHub; structured notebooks and clear documentation

View Projects Download CV Certifications Email Me

Outside work

I practice Brazilian Jiu-Jitsu (BJJ) and enjoy the mix of strategy, resilience, and continuous improvement—the same approach I bring to experiments and product polish.

Applying an armbar during a BJJ competition
Applying an armbar in competition.
Hand raised after a BJJ fight
A win—hand raised after a match.