I have a panda dataframe
with the following schema
:
customer_id int64
vehicle_type object
pickup_place object
place_category object
how_long_it_took_to_order object
pickup_lat float64
pickup_lon float64
dropoff_lat float64
dropoff_lon float64
pickup_coord object
dropoff_coord object
dtype: object
I am trying to find the distance between the pickup and drop locations. So I initially tried using the approach here Getting distance between two points based on latitude/longitude via the haversine formula. When i tried to conver the degrees to radians using
df_post['lat1'] = radians(df_post['pickup_lat'])
I got this error:
TypeError: cannot convert the series to <class 'float'>
So i tried following the approach in the 3rd reply, using the geopy.distance
module using the in-built function and for that created a tuple of the lat and long.
df_post['pickup_coord']=list(zip(df_post['pickup_lat'],df_post['pickup_lon']))
df_post['dropoff_coord']=list(zip(df_post['dropoff_lat'],df_post['dropoff_lon'])
But when i tried the in-built function
df_post['pickup_dropoff_distance']=gd.VincentyDistance(df_post['pickup_coord'],df_post['dropoff_coord']).miles
I am getting a new error:
ValueError: When creating a Point from sequence, it must not have more than 3 items.
Can someone help me with why either of the errors are ocurring and what is the possible solution.
The syntax for your distance calculator is geopy.distance.VincentyDistance(coords_1, coords_2).miles
, where coords_1
and coords_2
are tuples.
To apply the function to each row in a dataframe, you need to use pd.DataFrame.apply
:
def distancer(row):
coords_1 = (row['pickup_lat'], row['pickup_long'])
coords_2 = (row['dropoff_lat'], row['dropoff_long'])
return geopy.distance.VincentyDistance(coords_1, coords_2).miles
df_post['pickup_dropoff_distance'] = df_post.apply(distancer, axis=1)