boston_dataset_exploration

Dataset exploration: Boston house prices

The task is a basic analysis of the classic dataset Boston house prices.

Goals

• understand the dataset and problem associated with it
• examine the tools which help us describe and visualize the data
• make some artifacts - report of our findings, plots and code
• be prepared for the next dataset with the right questions and code

Questions

Inspired by the Applied Machine Learning Process book.

• Problem definition
  • What is the problem?
    • Informal description
    • Formal description
    • Assumptions
    • Similar problems
    • Description of provided data
  • Why does the problem need to be solved?
    • Motivation
    • Benefits
    • Use
  • How would I solve the problem (manually)?
• Data analysis
  • Summarize data
    • Data structure
    • Data distributions
      • Correlations with target variable
      • Correlations among attributes (redundancy)
  • Visualize data
    • Attribute histograms
    • Pairwise scatter-plots

Solution

See:

• problem_definition.md
• data_analysis.md
• data_analysis.py