# Datasheet for simular Datasets
## Dataset: California Cities TSP
### Motivation
**Purpose**: Benchmark dataset for traveling salesman problem (TSP) optimization algorithms.
**Creator**: simular development team
**Funding**: Open source project
### Composition
**Instances**: 25 California cities with real geographic coordinates
**Features per instance**:
- `name`: City name (string)
- `lat`: Latitude (float, degrees)
- `lon`: Longitude (float, degrees)
**Data format**: YAML
**Size**: ~2KB
**Sampling**: Hand-selected major California cities for diverse geographic distribution
### Collection Process
**Source**: Public geographic data (US Census, OpenStreetMap)
**Methodology**: Manual curation of 25 cities representing diverse California regions
**Timeframe**: December 2024
### Preprocessing
- Coordinates rounded to 4 decimal places
- Validated against multiple sources
- DVC tracked for version control
### Uses
**Intended use**: Benchmarking TSP optimization algorithms
**Not recommended**: Production routing applications (use actual road networks)
### Distribution
**License**: MIT (same as simular)
**Location**: `data/raw/california_cities.yaml`
### Maintenance
**Updates**: As needed for benchmark improvements
**Contact**: Via GitHub issues
---
## Dataset: Benchmark Metrics
### Composition
Three JSON files in `metrics/`:
- `tsp_benchmark.json`: TSP algorithm performance
- `orbit_benchmark.json`: Orbital mechanics simulation accuracy
- `monte_carlo_benchmark.json`: Monte Carlo estimation convergence
Each contains:
- Mean values with 95% confidence intervals
- Cohen's d effect sizes
- Sample sizes (n >= 100)
### License
MIT License
