# Kernel Density Estimation (KDE)

Generate density surfaces from point data using kernel density estimation.

## Overview

KDE creates a continuous density surface from point data, showing where points are concentrated. Higher values indicate greater point density.

## Inputs

- **Dataset**: Point dataset
- **Bandwidth**: Smoothing parameter (default: auto-calculated)
- **Cell Size**: Output raster cell size (default: auto-calculated)
- **Weight Field** (optional): Field to weight points

## Outputs

Raster dataset containing:

- Density values for each cell
- Higher values indicate greater point density
- Proper spatial reference

## Algorithm

1. Calculate optimal bandwidth (if not specified)
2. Create output raster grid
3. For each cell, calculate kernel-weighted sum of nearby points
4. Store density values in raster

## Example

```json
{
  "dataset_id": 123,
  "bandwidth": 1000,
  "cell_size": 100
}
```

## Background Jobs

This analysis runs as a background job.

## Use Cases

- Population density mapping
- Crime hotspot visualization
- Species distribution modeling
- Event density analysis

## Notes

- Bandwidth controls smoothing (larger = smoother)
- Cell size controls output resolution
- Weight field allows importance weighting
- Results are sensitive to bandwidth selection

## Related Documentation

- [Analysis API](../api/analysis.md)
- [Raster Tools](raster.md)