Geopandas: how to read a csv and convert to a geopandas dataframe with polygons?

Question

I read a .csv file as a dataframe that looks like the following:

import pandas as pd
df = pd.read_csv('myFile.csv')
df.head()
    BoroName    geometry
0   Brooklyn    MULTIPOLYGON (((-73.97604935657381 40.63127590...
1   Queens      MULTIPOLYGON (((-73.80379022888098 40.77561011...
2   Queens      MULTIPOLYGON (((-73.8610972440186 40.763664477...
3   Queens      MULTIPOLYGON (((-73.75725671509139 40.71813860...
4   Manhattan   MULTIPOLYGON (((-73.94607828674226 40.82126321...

I want to convert it to a geopandas dataframe.

import geopandas as gpd
crs = {'init': 'epsg:4326'}
gdf = gpd.GeoDataFrame(df, crs=crs).set_geometry('geometry')

but I get the following error

TypeError: Input must be valid geometry objects: MULTIPOLYGON (((-73.97604935657381 40.631275905646774, -73.97716511994669 40.63074665412933,....

ffrosch · Answer 1 · 2022-12-05T12:49:16.517

35

Geopandas seems to be unable to convert a geometry column from a pandas dataframe.

Solution number 2

Try applying the shapely wkt.loads function on your column before converting your dataframe to a geodataframe.

from shapely import wkt

df['geometry'] = df['geometry'].apply(wkt.loads)
gdf = gpd.GeoDataFrame(df, crs='epsg:4326')

Good luck!

Do not use - crashes spyder and jupyter kernel for some people

Solution number 1: Try loading the csv directly with geopandas

gdf = gpd.read_file('myFile.csv')
gdf.crs = 'epsg:4326'

edited Dec 05 '22 at 12:49

answered Apr 09 '20 at 14:03

ffrosch

899
6
15

4

number 2 works for me. number 1 kills my jupyter-notebook kernel for some reason – Sam Murphy Apr 10 '21 at 13:57
1

Number 1 also crashes Spyder kernel – Stelios K. Jun 11 '21 at 12:29
Number 2 works great – eeny Jan 14 '22 at 23:11

score 29 · Answer 2 · edited Mar 04 '22 at 12:46

29

You could also try this:

gdf = gpd.GeoDataFrame(
    df, geometry=gpd.points_from_xy(df.longitude, df.latitude)
)

This will convert those lat/long columns to points

edited Mar 04 '22 at 12:46

James

32,991
4
47
70

answered Jun 28 '20 at 01:03

SKhan2312

297
3
3

1

But where within the dataframe presented are the columns latitude and longitude? – shadow_dev Jul 09 '20 at 21:43
1

The correct order when using geometric coordinates with `points_from_xy` is `points_from_xy(longitudes, latitudes)`. – mins Jan 04 '22 at 14:25

score 1 · Answer 3 · answered Jun 09 '22 at 15:59

1

Dudes, there are wkt geometry strings in the original dataframe, not xy columns, so i would suggest read this: DataFrame with WKT Column to GeoPandas Geometry

answered Jun 09 '22 at 15:59

sal

1,199
1
13
31

score 0 · Answer 4 · answered Nov 17 '22 at 05:00

0

Geopandas puts a geometry column at the end if you load directly. Found this out by experimenting with column names and it worked

answered Nov 17 '22 at 05:00

gillan_data

1
1

Geopandas: how to read a csv and convert to a geopandas dataframe with polygons?

4 Answers4

Linked