Understanding Geography Types

What is Geography in PostGIS?

PostGIS provides two primary types to store spatial data: geometry and geography. While geometry uses a flat Cartesian plane, geography understands the Earth as a sphere. This makes geography perfect for global-scale data where curvature matters.

If you're working on distances between cities, flight paths, or cross-country analysis, geography is your best choice for accurate results.

Geometry vs Geography – Key Differences

Rule of thumb: For city zoning, use geometry. For airline distance between Mumbai and New York, use geography.

Creating Geography Columns

Creating tables with geography columns is straightforward:

CREATE TABLE capitals (
  id SERIAL PRIMARY KEY,
  name TEXT,
  country TEXT,
  location GEOGRAPHY(Point, 4326)
);

INSERT INTO capitals (name, country, location) VALUES
('Delhi', 'India', ST_GeogFromText('POINT(77.2090 28.6139)')),
('Paris', 'France', ST_GeogFromText('POINT(2.3522 48.8566)')),
('Tokyo', 'Japan', ST_GeogFromText('POINT(139.6917 35.6895)')),
('New York', 'USA', ST_GeogFromText('POINT(-74.0060 40.7128)')),
('Sydney', 'Australia', ST_GeogFromText('POINT(151.2093 -33.8688)'));

Geography Input Functions

Distance Calculations with Geography

One of the primary advantages of geography is accurate distance calculation:

-- Distance in meters between two capitals
SELECT
  a.name AS from_city,
  b.name AS to_city,
  ROUND(ST_Distance(a.location, b.location)) AS distance_meters,
  ROUND(ST_Distance(a.location, b.location) / 1000) AS distance_km
FROM capitals a, capitals b
WHERE a.name = 'Delhi' AND b.name = 'Paris';

Proximity Searches

Use ST_DWithin for proximity searches with geography:

-- Find capitals within 1000 km of Delhi
SELECT name, country
FROM capitals
WHERE ST_DWithin(
  location, 
  ST_GeogFromText('POINT(77.2090 28.6139)'), 
  1000000  -- 1000 km in meters
)
AND name != 'Delhi';

-- Find the closest capital to a given point
SELECT name, country, 
       ROUND(ST_Distance(location, ST_GeogFromText('POINT(0 51.5)'))) AS distance_m
FROM capitals
ORDER BY location <-> ST_GeogFromText('POINT(0 51.5)')
LIMIT 1;

Geography Types Supported

Working with Larger Geographic Features

-- Create a table for countries
CREATE TABLE countries (
  id SERIAL PRIMARY KEY,
  name TEXT,
  boundary GEOGRAPHY(MultiPolygon, 4326)
);

-- Example flight routes
CREATE TABLE flight_routes (
  id SERIAL PRIMARY KEY,
  airline TEXT,
  route_name TEXT,
  path GEOGRAPHY(LineString, 4326)
);

INSERT INTO flight_routes (airline, route_name, path) VALUES
('Air India', 'Delhi-London', 
 ST_GeogFromText('LINESTRING(77.2090 28.6139, -0.1276 51.5074)'));

-- Calculate flight distance
SELECT 
  route_name,
  ROUND(ST_Length(path) / 1000) AS distance_km
FROM flight_routes;

Spatial Catalog for Geography

PostGIS tracks geography columns in a dedicated catalog table:

-- View all geography columns in the database
SELECT 
  f_table_schema,
  f_table_name,
  f_geography_column,
  type,
  srid,
  coord_dimension
FROM geography_columns;

-- Get detailed information about a specific geography column
SELECT 
  f_table_name AS table_name,
  f_geography_column AS column_name,
  type AS geography_type,
  coord_dimension AS dimensions
FROM geography_columns
WHERE f_table_name = 'capitals';

Performance Considerations

Indexing Geography Columns

-- Create spatial index (automatically uses GIST)
CREATE INDEX idx_capitals_location ON capitals USING GIST(location);

-- Check index usage
EXPLAIN ANALYZE
SELECT name FROM capitals
WHERE ST_DWithin(location, ST_GeogFromText('POINT(77.21 28.61)'), 500000);

Performance Comparison

Limitations and Considerations

Geography Limitations

• Function support: Geography supports fewer functions than geometry (no TINs, curves, or complex topology operations)
• Performance: Spatial joins and complex operations may be slower
• Projection: Cannot use arbitrary projections—only WGS84 (SRID 4326)
• Precision: Limited to spherical calculations (no ellipsoidal refinements for highest precision)

When to Choose Geography vs Geometry

Real-World Applications

Global Logistics

-- Calculate shipping distances between ports
CREATE TABLE ports (
  name TEXT,
  country TEXT,
  location GEOGRAPHY(Point, 4326)
);

-- Find optimal shipping routes
SELECT 
  p1.name AS origin,
  p2.name AS destination,
  ROUND(ST_Distance(p1.location, p2.location) / 1852) AS nautical_miles
FROM ports p1, ports p2
WHERE p1.name != p2.name
ORDER BY ST_Distance(p1.location, p2.location)
LIMIT 10;

Climate Analysis

-- Analyze weather stations within climate zones
CREATE TABLE weather_stations (
  id SERIAL PRIMARY KEY,
  name TEXT,
  location GEOGRAPHY(Point, 4326),
  elevation INTEGER
);

-- Find stations within 500km of a reference point
SELECT name, elevation
FROM weather_stations
WHERE ST_DWithin(
  location,
  ST_GeogFromText('POINT(77.21 28.61)'),
  500000
);

Best Practices

1. Evaluate your scale: Use geography for global/national analysis, geometry for local projects
2. Index geography columns: Always create spatial indexes for better performance
3. Validate coordinates: Ensure longitude is between -180 and 180, latitude between -90 and 90
4. Consider precision needs: Geography provides good accuracy for most applications
5. Test performance: Compare query performance between geography and projected geometry