EDA & Machine Learning exercises

• ❤️‍🔥 OKcupid

  • Dataset : OKcupid Profiles
  • Mini EDA exercise : OKcupid EDA
  • User segmentation : OKcupid Customer segmentation
    • Lessons learned : The concept of clustering is more or less clear enough to apply. However, I am still having difficulties in understanding how to interpret the result of clustering algorithms 🤷🏻‍♀️ Higher number of featrues seems to make more difficult to cluster datapoints and visualize the result.

• 🚴🏻‍♂️ Stolen bicycle report in Berlin Germany

  • Dataset : https://www.kaggle.com/datasets/staefff/stolen-bike-report-berlin
  • EDA exercise: Berlin stolen bicycle report
  • Lessons learned: shame that I couldnt do forecasting because there were many missing dates within the period. One year dataset will lead to overfitting with high possibility.
  • Tableau visualization:
  • Shared in Kaggle: Kaggle

• 🐺 Woes Club Churn Prediction with Pycaret

  • Dataset: Woes Club
  • References: Pycaret
  • Churn Prediction Model
  • Lesson learned: Pycaret needs certain environment settings. It is definitely a useful package to easily get the best model for dataset. But I would not totally rely on the package. Always try out different approaches for data preprocessing and model building.

• 🚢 E-Commerce Lieferverzug-Prognose

  • Dataset: Product Shipment Delivered on time or not? To Meet E-Commerce Customer Demand
  • Lieferverzug-Prognose