contextily making weird background maps

Question

This is my code:

import pandas as pd
import geoplot as gplt
import geopandas as gpd
import geoplot.crs as gcrs
import contextily

df = pd.read_csv('dataframe_master.csv', index_col='id')

crs = {'init': 'epsg:4326'}
geometry = [geometry.Point(xy) for xy in zip(df['latitude'], df['longitude'])]
df_geo = gpd.GeoDataFrame(df_geo, crs=crs, geometry=geometry)

test = df_geo[:200000]
test = test.to_crs(epsg=3857)

ax = test.plot(marker='o', markersize=1)
contextily.add_basemap(ax)
plt.show()

And it generates this image: image, which doesn't show a background map and seems a little distorted.

My coordinate data was originally made with the RD-coordinaten standard (EPSG:28992), which I converted to EPSG:4326 with this code:

lon_l = []
lat_l = []

p1 = Proj(init='epsg:28992')
p2 = Proj(proj='latlong',datum='WGS84')
 
for row in range(len(df)):
    lon, lat, z = transform(p1, p2, df.iloc[row, 7], df.iloc[row, 8], 0.0)
    lon_l.append(lon)
    lat_l.append(lat) 

I did a sanity check on the longitude latitude output by comparing to some online converters, and the output points to the correct locations.

I tried following this solution: https://gis.stackexchange.com/questions/348339/using-crs-epsg3857-but-misalignment-between-stamen-background-and-coordinates-o in case my conversion was missing the "towgs84"part, but the image still looked the same with a slightly different colour.

Answer 1

I figured it out! I should've listed longitude before latitude when building the geometry.

geometry = [geometry.Point(xy) for xy in zip(df['longitude'], df['latitude'])]