CODEWITHSUNDEEP
pythonpython-3.xgeopandas
M.O.
M.O.

Reputation: 496

Geopandas - ValueError: Cannot transform naive geometries. Please set a crs on the object first

Can someone please explain how this is not working?

enter image description here

The shapefile in question is here and in the code its read as

shp=gpd.read_file("Microdatos_Censo_2017_Manzana/Microdatos_Censo_2017_Manzana.shp")
shp.crs="epsg:4326"
breakpoint()
shp=shp.to_crs(epsg=3857)## Error here

I just dont get what is happening. I have Python 3.8.5, geopandas 0.8.1, pyproj 2.6.1.post1. Not sure what other package would be important to know versions of.

Thanks!

Edit:

1.- Fixing link to shapefile that I had gotten wrong.

2.- It's not the same as the question that is posted as a duplicate, because as you can see on the image the print statement of shp.crs returns the correct crs information, not None. I have a crs defined and yet to_crs is not working.

Upvotes: 7

Views: 13471

Answers (2)

Emad Armiti
Emad Armiti

Reputation: 1

In my case I was trying to project LineStrings, it didn't work. I had to first project Points then create LineStrings.

Upvotes: 0

swatchai
swatchai

Reputation: 18762

In your code:

shp = gpd.read_file("Comunas/Comunas.shp")

let you get shp as a GeoDataFrame.

Next, the line

shp.crs = "epsg:4326"

only changes the property of shp, but does not perform coordinate transformation on the geodataframe.

Then

shp = shp.to_crs(epsg=3857)

causes the error.

From the error messages it is obvious that the command that causes error requires proper object as its input value. The value in epsq=3857 is wrong according to the method's signature as follows.

.to_crs(crs=None, epsg=None, inplace=False)

The proper use of this method can be:

.to_crs({'init': 'epsg:4326'})
.to_crs(crs={'init': 'epsg:4326'})
.to_crs(epsg='4326')

For your particular dataset, to convert the CRS of the original GeoDataFrame (epsg:3857), to epsg:4326, and back to original, do these steps:

shp_file = './data/comunas/comunas.shp'  #(on my machine)
comunas0 = gpd.read_file(shp_file)
print(comunas0.crs)                      #{'init': 'epsg:3857'}
comunas0.plot()

epsg3857 (image of comunas0)

comunas4326 = comunas0.to_crs({'init': 'epsg:4326'})
print(comunas4326.crs)                   #{'init': 'epsg:4326'}
comunas4326.plot()

epsg4326 (image of comunas4326)

comunas3857 = comunas4326.to_crs(epsg='3857')  #back to original CRS
print(comunas3857.crs)                   #{'init': 'epsg:3857', 'no_defs': True}

EDIT

Additional plots using new shapefiles (as updated by OP).

epsg:3857

new3857

epsg:4326

new4326

Upvotes: 5

Related Questions