28

I have polygons in WKT that looks like this

POLYGON ((24.8085317 46.8512821, 24.7986952 46.8574619, 24.8088238 46.8664741, 24.8155239 46.8576335, 24.8085317 46.8512821))

I would like to calculate the area of this polygon in km². What would be the best way to do this in Python?

2
  • 2
    You may look at stackoverflow.com/questions/23697374/…
    – SIslam
    Dec 26 '14 at 9:12
  • If you're getting the following error implementing one of the above solutions, is because the lat1 and lat2 should be lat_1 and lat_2: pyproj.exceptions.CRSError: Invalid projection: +proj=aea +lat1=37.843975868971484 +lat2=37.844325658890924 +type=crs: (Internal Proj Error: proj_create: Error -21: conic lat_1 = -lat_2) Apr 17 '19 at 0:18
30

It wasn't readily apparent to me how to use @sgillies answer, so here is a more verbose version:

import pyproj    
import shapely
import shapely.ops as ops
from shapely.geometry.polygon import Polygon
from functools import partial


geom = Polygon([(0, 0), (0, 10), (10, 10), (10, 0), (0, 0)])
geom_area = ops.transform(
    partial(
        pyproj.transform,
        pyproj.Proj(init='EPSG:4326'),
        pyproj.Proj(
            proj='aea',
            lat_1=geom.bounds[1],
            lat_2=geom.bounds[3]
        )
    ),
    geom)

# Print the area in m^2
print(geom_area.area)
3
  • 2
    The resulting value is not exactly the same as the one obtained in geojson.io. Why? Jul 6 '17 at 19:58
  • What exactly did you do there? Oct 5 at 19:31
  • Any thoughts of solving this future warning: "crs.py:131: FutureWarning: '+init=<authority>:<code>' syntax is deprecated. '<authority>:<code>' is the preferred initialization method. When making the change, be mindful of axis order changes: pyproj4.github.io/pyproj/stable/… in_crs_string = _prepare_from_proj_string(in_crs_string)" ?
    – gcamargo
    Nov 3 at 16:30
19

It looks like your coordinates are longitude and latitude, yes? Use Shapely's shapely.ops.transform function to transform the polygon to projected equal area coordinates and then take the area.

python
import pyproj
from functools import partial

geom_aea = transform(
partial(
    pyproj.transform,
    pyproj.Proj(init='EPSG:4326'),
    pyproj.Proj(
        proj='aea',
        lat1=geom.bounds[1],
        lat2=geom.bounds[3])),
geom)

print(geom_aea.area)
# Output in m^2: 1083461.9234313113 
3
  • 1
    You should probably indicate that partial is not a built-in; pyproj will have to be imported and possibly installed, etc.
    – kingledion
    Apr 26 '18 at 18:27
  • I noticed pyproj.Proj(proj='aea', lat1=lat1, lat2=lat2) resulted in CRSError: Invalid projection: +proj=aea +lat1=5.0 +lat2=6.0 +type=crs. Changing lat{1,2} into lat_{1,2} as implied by proj4 documentation fixed it: pyproj.Proj(proj='aea', lat1=lat1, lat2=lat2). Is this Answer accurate, or should it be updated?
    – Herbert
    Mar 18 '19 at 15:00
  • 1
    I needed to use lat_1 and lat_2 instead of lat1 and lat2. I suspect this applies post PROJ 6.0.0
    – oortCloud
    Apr 26 '19 at 1:58
10

The above answers seem to be correct, EXCEPT that at some point recently, the lat1 and lat2 parameters in the pyproj code were renamed with underscores: lat_1 and lat_2 (see https://stackoverflow.com/a/55259718/1538758). I don't have enough rep to comment, so I'm making a new answer (sorry not sorry)

import pyproj    
import shapely
import shapely.ops as ops
from shapely.geometry.polygon import Polygon
from functools import partial


geom = Polygon([(0, 0), (0, 10), (10, 10), (10, 0), (0, 0)])
geom_area = ops.transform(
    partial(
        pyproj.transform,
        pyproj.Proj(init='EPSG:4326'),
        pyproj.Proj(
            proj='aea',
            lat_1=geom.bounds[1],
            lat_2=geom.bounds[3])),
    geom)

# Print the area in m^2
print geom_area.area 
1
  • Nice answer which seems to work for me but it performs "only" at 25 Iterations/second.
    – Stücke
    Oct 5 at 11:51
7

Calculate a geodesic area, which is very accurate and only requires an ellipsoid (not a projection). This can be done with pyproj 2.3.0 or later.

from pyproj import Geod
from shapely import wkt

# specify a named ellipsoid
geod = Geod(ellps="WGS84")

poly = wkt.loads('''\
POLYGON ((24.8085317 46.8512821, 24.7986952 46.8574619,
24.8088238 46.8664741, 24.8155239 46.8576335, 24.8085317 46.8512821))''')

area = abs(geod.geometry_area_perimeter(poly)[0])

print('# Geodesic area: {:12.3f} m²'.format(area))
print('#                {:12.3f} km²'.format(area/1e6))

# Geodesic area:  1083466.869 m²
#                       1.083 km²

abs() is used to return only positive areas. A negative area may be returned depending on the winding direction of the polygon.

5

I've stumbled across "area" which seems simpler to use:

https://pypi.org/project/area/

For example:

from area import area

obj = {'type':'Polygon','coordinates':[[[24.8085317,46.8512821], [24.7986952,46.8574619], [24.8088238,46.8664741], [24.8155239,46.8576335], [24.8085317,46.8512821]]]}

area_m2 = area(obj)

area_km2 = area_m2 / 1e+6
print ('area m2:' + str(area_m2))
print ('area km2:' + str(area_km2))

... returns:

area m2:1082979.880942425

area km2:1.082979880942425

2

I've found that the projected option is a bit too slow for my use so I did some digging and found a javascript lib from Mapbox called cheap-ruler. The "cheap" in the name is intended to imply that this algorithm isn't intended to return the right answer, but simply a good enough one.

Below is my copy and paste port of the area function to work with a shapely geom. My port doesn't remove area for holes, but it could be updated to do so.

import math

def area(geom):
    ## Setup from:
    ## https://github.com/mapbox/cheap-ruler/blob/48ad4768a52dc176b01494d090cce19f02c7afdd/index.js#L71-L82
    RE = 6378.137
    RAD = math.pi / 180
    FE = 1 / 298.257223563
    E2 = FE * (2 - FE)

    lat = geom.centroid.y
    m = RAD * RE * 1000
    coslat = math.cos(lat * RAD);

    w2 = 1 / (1 - E2 * (1 - coslat * coslat))
    w = math.sqrt(w2)
    kx = m * w * coslat
    ky = m * w * w2 * (1 - E2)

    ## Area calc from:
    ## https://github.com/mapbox/cheap-ruler/blob/48ad4768a52dc176b01494d090cce19f02c7afdd/index.js#L185-L197
    ring = geom.exterior.coords

    sumVal = 0
    j = 0
    l = len(ring)
    k = l - 1
    while j < l:
        sumVal += wrap(ring[j][0] - ring[k][0]) * (ring[j][1] + ring[k][1])
        k = j
        j += 1

    return (abs(sumVal) / 2) * kx * ky;

def wrap(deg):
    while deg < -180:
        deg += 360
    while deg > 180:
        deg -= 360
    return deg
2
2

As you probably use this in a loop for many geometries you can get faster transformations by using modern Proj as intended and by not recreating static objects like the initial CRS:

from pyproj import Transformer, transform, CRS
from shapely.geometry import Polygon
import shapely.ops as ops

geom = Polygon([(0, 0), (0, 10), (10, 10), (10, 0), (0, 0)])

crs_4326 = CRS.from_epsg(4326)

transformer = Transformer.from_crs(
    crs_4326,
    CRS(proj='aea', 
        lat_1=geom.bounds[1],
        lat_2=geom.bounds[3]
    )
)

geom_area = ops.transform(transformer.transform, geom)

is about twice as fast for me if you measure the setup of the transformer and the transformation.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.