Plotting Lat/Long Points Using Basemap

I am trying to plot points on a map using matplotlib and Basemap, where the points represent the lat/long for specific buildings. My map does indeed plot the points, but puts them in the wrong location. When I use the same data and do the same thing using Bokeh, instead of matplotlib and basemap, I get the correct plot.

Here is the CORRECT result in Bokeh: Bokeh Version

And here is the INCORRECT result in Basemap: Basemap Version

I have seen discussion elsewhere on StackOverflow that suggested this might be related to the fact that plot() "shifts" the longitude somehow. I've tried the suggestion from there, which was to include the line: lons, lats = m.shiftdata(long, lat) and then use the shifted data. That didn't have any visible impact.

My full sample code which generates both of the plots in Basemap and Bokeh is here:

import matplotlib.pyplot as plt
from mpl_toolkits.basemap import Basemap
import pandas as pd

from bokeh.plotting import figure, show
from bokeh.sampledata.us_states import data as states
from bokeh.models import ColumnDataSource, Range1d

# read in data to use for plotted points
buildingdf = pd.read_csv('buildingdata.csv')
lat = buildingdf['latitude'].values
long = buildingdf['longitude'].values

# determine range to print based on min, max lat and long of the data
margin = .2 # buffer to add to the range
lat_min = min(lat) - margin
lat_max = max(lat) + margin
long_min = min(long) - margin
long_max = max(long) + margin

# create map using BASEMAP
m = Basemap(llcrnrlon=long_min,
            llcrnrlat=lat_min,
            urcrnrlon=long_max,
            urcrnrlat=lat_max,
            lat_0=(lat_max - lat_min)/2,
            lon_0=(long_max-long_min)/2,
            projection='merc',
            resolution = 'h',
            area_thresh=10000.,
            )
m.drawcoastlines()
m.drawcountries()
m.drawstates()
m.drawmapboundary(fill_color='#46bcec')
m.fillcontinents(color = 'white',lake_color='#46bcec')
# convert lat and long to map projection coordinates
lons, lats = m(long, lat)
# plot points as red dots
m.scatter(lons, lats, marker = 'o', color='r')
plt.show()


# create map using Bokeh
source = ColumnDataSource(data = dict(lat = lat,lon = long))
# get state boundaries
state_lats = [states[code]["lats"] for code in states]
state_longs = [states[code]["lons"] for code in states]

p = figure(
           toolbar_location="left",
           plot_width=1100,
           plot_height=700,
           )

# limit the view to the min and max of the building data
p.y_range = Range1d(lat_min, lat_max)
p.x_range = Range1d(long_min, long_max)
p.xaxis.visible = False
p.yaxis.visible = False
p.xgrid.grid_line_color = None
p.ygrid.grid_line_color = None

p.patches(state_longs, state_lats, fill_alpha=0.0,
      line_color="black", line_width=2, line_alpha=0.3)

p.circle(x="lon", y="lat", source = source, size=4.5,
         fill_color='red',
         line_color='grey',
         line_alpha=.25
         )
show(p)

I don't have enough reputation points to post a link to the data or to include it here.

In the basemap plot the scatter points are hidden behind the fillcontinents. Removing the two lines

#m.drawmapboundary(fill_color='#46bcec')
#m.fillcontinents(color = 'white',lake_color='#46bcec')

would show you the points. Because this might be undesired, the best solution would be to place the scatter on top of the rest of the map by using the zorder argument.

m.scatter(lons, lats, marker = 'o', color='r', zorder=5)

Here is the complete code (and I would like to ask you to include this kind of runnable minimal example with hardcoded data next time asking a question, as it saves everyone a lot of work inventing the data oneself):

import matplotlib.pyplot as plt
from mpl_toolkits.basemap import Basemap
import pandas as pd
import io

u = u"""latitude,longitude
42.357778,-71.059444
39.952222,-75.163889
25.787778,-80.224167
30.267222, -97.763889"""

# read in data to use for plotted points
buildingdf = pd.read_csv(io.StringIO(u), delimiter=",")
lat = buildingdf['latitude'].values
lon = buildingdf['longitude'].values

# determine range to print based on min, max lat and lon of the data
margin = 2 # buffer to add to the range
lat_min = min(lat) - margin
lat_max = max(lat) + margin
lon_min = min(lon) - margin
lon_max = max(lon) + margin

# create map using BASEMAP
m = Basemap(llcrnrlon=lon_min,
            llcrnrlat=lat_min,
            urcrnrlon=lon_max,
            urcrnrlat=lat_max,
            lat_0=(lat_max - lat_min)/2,
            lon_0=(lon_max-lon_min)/2,
            projection='merc',
            resolution = 'h',
            area_thresh=10000.,
            )
m.drawcoastlines()
m.drawcountries()
m.drawstates()
m.drawmapboundary(fill_color='#46bcec')
m.fillcontinents(color = 'white',lake_color='#46bcec')
# convert lat and lon to map projection coordinates
lons, lats = m(lon, lat)
# plot points as red dots
m.scatter(lons, lats, marker = 'o', color='r', zorder=5)
plt.show()