add_station_lat_lon yields columns with object dtype #2872
Description
df = pd.DataFrame({'station': ['KOUN', 'KVPZ', 'KDEN', 'PAAA']})
df = add_station_lat_lon(df, 'station')
df.info()gives
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 4 entries, 0 to 3
Data columns (total 3 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 station 4 non-null object
1 latitude 3 non-null object
2 longitude 3 non-null object
dtypes: object(3)
memory usage: 224.0+ bytes
Apparently it's good enough to pass e.g. assert_almost_equal(df.loc[df.station == 'KOUN'].latitude.values[0], 35.25), but it's painful to pass to e.g. tricontour().