4

I have a geodataframe that has no assigned crs. I want to create a column/field called 'area' that has the area calculated as square feet, not meters or km, using geopandas. Note: this geodataframe was created by doing an attribute join between 2 previous geodataframes that had the projection: +proj=longlat +datum=WGS84 +no_defs 

[38]: print(bldgs_merged.crs)
      None

Something like this:

bldgs_merged["bldg_sqft"] = bldgs_merged['geometry'].area/ 10**6

I know that the crs/projection first needs to be in a crs/projection that uses square feet. My question is, what projection do I set it to? Will WGS 84/EPSG 4326 work? Do I have to do something to set the units to be square feet?

bldgs_merged = bldgs_merged.to_crs({'init': 'epsg:4326'})
gwydion93
  • 1,903
  • 4
  • 28
  • 59

1 Answers1

2

the area attribute of a GeoDataFrame is calculated based on projected coordinates not latitude and logitude, this means that the area you are calculating is false.

First you need to reproject the GeoDataFrame to a different CRS which units is either meter or feet, and then measure the area using areaattribute.

PS: EPSG:4326 is using degrees as units, try the EPSG:3857 which uses meters as unit or if you know the location of your polygons choose a projection from a national grid

Hicham Zouarhi
  • 3,255
  • 2
  • 19
  • 32
  • Question. This seemed to work, but the calculation of the area from the geometry is returned as dtype: float with a nonsensical value like 7.338585e-09. Does this look right? How would I convert this to a 2-digit double? – gwydion93 Jul 19 '19 at 16:44
  • 1
    @gwydion93 you can round the result to 2 digit by using a lambda function gdf['rounded_area'] = gdf['area'].apply(lambda x: round(x, 2)) – Hicham Zouarhi Jul 19 '19 at 20:27