Freelance data scientist · All case studies

Who buys the next triathlon watch.

A full preprocessing track, recursive feature elimination, then a model bake-off — decision trees, MLP, SVM, logistic regression — to map early-adopter profiles for an upcoming smartwatch.

Client: Wearable launch study. Built by Dilshad Raza.

Early adopters for a triathlon watch sit in a thin slice of the customer file. Dummy variables, scale, and a kitchen-sink feature dump will happily overfit that slice. The work had to preprocess cleanly, drop weak indicators, and compare learners on the same pipe.

I implemented the full preprocessing track, used recursive feature elimination to keep the indicators that actually moved early-adopter status, and trained decision trees, an MLP classifier, SVCs, and logistic regression so the launch team could see stability, not just a favourite AUC.

Hire Dilshad Raza for similar freelance data science, machine learning, and AI automation in the UK, United States, and Australia.