Histogram
Box Plot
Violin Plot
Distribution Type
Sample Size
Visualization
0
Mean
0
Median
0
Std Dev
0
Skewness
0
Kurtosis
0
Range
Understanding Data Distributions
Understanding the distribution of your data is fundamental to data science. It guides algorithm choice, feature engineering, and helps detect data quality issues.
Key Statistical Measures
- Mean: Average value, sensitive to outliers
- Median: Middle value, robust to outliers
- Mode: Most frequent value
- Standard Deviation: Spread of data around mean
- Variance: Square of standard deviation
- Skewness: Asymmetry of distribution
- Kurtosis: "Tailedness" of distribution
Common Distributions
- Normal (Gaussian): Bell curve, many natural phenomena
- Uniform: All values equally likely
- Skewed: Long tail on one side
- Bimodal: Two peaks, mixed populations
- Exponential: Waiting times, decay processes
- Power Law: Heavy-tailed, many real-world networks
Why Distribution Matters for ML
- Many algorithms assume normality (linear regression, LDA)
- Skewed features may need transformation
- Outliers affect training
- Scaling methods depend on distribution
- Distribution shifts indicate data drift
Visualization Types
- Histogram: Shows frequency distribution
- Box Plot: Shows quartiles and outliers
- Violin Plot: Combines box plot with density
- QQ Plot: Compare against theoretical distribution
- ECDF: Empirical cumulative distribution
Data Transformations
- Log Transform: For right-skewed data
- Square Root: Moderate skewness
- Box-Cox: Optimal power transformation
- Yeo-Johnson: Handles negative values
- Quantile Transform: Force to normal distribution
Experiment with the Explorer
Use the interactive tool above to:
- Generate different distributions
- Compare visualization types
- See how sample size affects appearance
- Understand statistical measures
- Identify outliers and anomalies