Data Exploration

• Summarisation
  Start by determining basic insights about the data

  • Mean

    • Avoided when large outliers are present in the data

  • Median

  • Mode

  • IQR

  • Standard Deviation

  • Class Imbalance

    • If one class dominates or a specific class has a low count

• Visualisation
  Select a subset of data and choose an appropriate graph to visualise it

  • Histograms

    • For Numerical Data

  • Bar Charts

    • For Categorical Data

  • Scatter Plots

    • To determine correlation and outliers

  • Line Plots

    • Connected relation between data

  • Box Plots

    • To determine the distribution, outliers and quartiles of a numerical dimensions

  • Contour Plots

    • Used for temporal Data

  • Correlation Matrix