2

I have two GeoDataframes - one has polygons, the other points. 

import geopandas as gpd
from shapely.geometry import Point
area = gpd.GeoDataFrame.from_file('areas.shp')
print area.crs

{u'lon_0': -91.8666666667, u'datum': u'NAD83', u'y_0': 3000000, u'no_defs': True, u'proj': u'lcc', u'x_0': 6200000, u'units': u'm', u'lat_2': 77, u'lat_1': 49, u'lat_0': 63.390675}

Addresses were assigned geolocations through an API, so I assigned a CRS 4326, as that's the one most commonly used it seems.

location['geometry'] = location.Geolocation.apply(lambda x: Point(eval(x)))
gdf = gpd.GeoDataFrame(location, geometry='geometry', crs={'init': 'epsg:4326'})

Then I changed the area's CRS:

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

And tried to merge the two dataframes

merged = gpd.sjoin(area, gdf, op='intersects')

But the result it gives me is Empty GeoDataFrame. I am not sure why, because the points definitely do fall inside polygons. What am I doing wrong?

Here are my files

https://www.dropbox.com/sh/9lioawgsor5y4wc/AABkaaFpjAjisQEVaLrFwSnsa?dl=0

I suspect that its because of the points that were somehow not properly processed into the right coordinate system maybe.

Improve this question
4
  • 1
    Have you tried plotting them to make sure they overlap? geopandas.org/mapping.html#maps-with-layers Commented Jul 21, 2018 at 7:19
  • If you can post some sample data perhaps? Commented Jul 21, 2018 at 23:50
  • @BERA, I have, the points for some reason are not plotting right, that probably is my problem there. Commented Jul 25, 2018 at 3:29
  • 1
    @om_henners, I just added them to the original question Commented Jul 25, 2018 at 3:29

1 Answer 1

Reset to default
4

Looking at your data, the issue is in the property points that you've geocoded:

|                     Property Address |               Geolocation |                      geometry |
|--------------------------------------|---------------------------|-------------------------------|
| 1586 LORNE ST E,  Kamloops BC Canada | (50.673675,  -120.298973) | POINT (50.673675 -120.298973) |

When you've geocoded the addresses you've got latitude and longitude, which is fine, but you've created the points in that order. Latitude and Longitude are Y, X measures respectively. You'll have to swap the coordinates when you create the points in your Geolocation (and doing so works fine in terms of the spatial join).

Improve this answer
1
  • Thank you for spotting that, I would've never guessed that is where the mistake was. But now I'm facing a problem with swapping the coordinates. Below is the script I wrote that is for some reason not functioning. for u in gdf['Geolocation']: u = u.replace('(', '') u = u.replace(')', '') a, b = u.split(',') correct = [b, a] correct = gdf['Geolocation_correct'] Commented Aug 6, 2018 at 6:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.