Projects

Six hands-on projects applying high-dimensional data analysis to real public datasets, one per sector. Each has a code repository, a plain-language write-up and a short theory note.

• MarketingPlanned

  Customer segmentation from retail transactions

  Group customers into actionable segments using PCA and clustering.

• IndustryPlanned

  Fault detection from semiconductor sensors

  Flag faulty production runs from about 590 sensors with very few failures.

• FinancePlanned

  Portfolio construction with covariance shrinkage

  Compare raw and shrunk covariance estimates in an out-of-sample backtest.

• HealthcarePlanned

  Gene selection with more features than samples

  Select informative genes from thousands of features using regularised regression.

• Text and AIPlanned

  Semantic search and the cost of dimension reduction

  Build a vector search tool and measure accuracy lost as dimensions shrink.

• EducationPlanned

  Automated answer evaluation with uncertainty

  Score written answers from embeddings and compare to human scores with confidence intervals.