Distribution Analysis identifies the distribution type for each numeric column and provides key statistical measures.
Columns:
Column: Name of the numeric column
Distribution: Inferred distribution type (Normal, LogNormal, Uniform, PowerLaw, Exponential)
Shapiro-Wilk: W statistic from Shapiro-Wilk normality test (0-1, higher = more normal)
SW p-value: P-value from Shapiro-Wilk test (probability of observing W under normality)
CV: Coefficient of variation (std/mean, relative variability independent of scale)
Outliers: Count and percentage of outliers (IQR method)
Skewness: Asymmetry measure (positive = right-tailed, negative = left-tailed)
Kurtosis: Tail heaviness compared to normal distribution (3.0 = normal)
Color Coding:
Distribution types are color-coded:
- Green/Cyan: Good fit quality (>0.75)
- Yellow: Moderate fit quality (≤0.75)
- Red: Very high outlier percentage (>20%) or extreme skewness/kurtosis
Navigation:
↑↓ / j/k: Navigate rows
←→ / h/l: Scroll columns horizontally
Tab: Switch focus between main area and sidebar
Enter: Open detail view for selected column (shows Q-Q plot and histogram)
Esc: Close analysis view
r: Resample data (only shown if data was sampled)
Detail View:
Press Enter on a row to see detailed analysis with Q-Q plots and histograms comparing your data to theoretical distributions.