How to find nearest distance from a coordinate point inside a polygon to the edge (border) of the polygon in Python?

I currently have a dataset with lat and lon coordinates in my first dataframe.

And then I have a shape of a country, Nigeria.

world_filepath = gpd.datasets.get_path('naturalearth_lowres')

world = gpd.read_file(world_filepath)

nigeria = world.loc[world['name'] == 'Nigeria']

pop_est      continent	name	iso_a3	gdp_md_est	geometry

56 200963599.0 | Africa | Nigeria NGA | 448120 | POLYGON ((26958.294 253210.950, 35554.540 4322..

I am not sure how to find the NEAREST distance between the coord points in my first dataframe to the nigeria shape border. I have tried using shapely i think but I keep getting 0 since the coordinates are inside of the polygon.

At the end result, I want to the distance from the coord point to the nearest distance to the boarder of Nigeria for all the rows in the first dataframe.

Thank you

Edit -- To include more info and code
I saved the first dataset with the point coordinates as 'coord_new'

df=gpd.GeoDataFrame(coord_new,geometry=gpd.points_from_xy(coord_new.longitude, coord_new.latitude))

#df= df.to_crs(26392)

#nigeria = nigeria.to_crs(26392)

df['distance1']=nigeria.exterior.distance(df)

df['distance1']=nigeria.boundaries.distance(df)

This is the way that I saw other people doing to find the distance from point to the border of the polygon. However, I keep getting 0 and then NaN in the distance1 column.

The issue you're encountering is that when using binary spatial operations or distance involving two GeoDataFrames, the DataFrames are aligned on their indexes.

Note the arguments to the GeoSeries.distance function:

> Parameters:
> * other: Geoseries or geometric object
> The Geoseries (elementwise) or geometric object to find the distance to.

When you subset the global shapefile to only rows matching the name "Nigeria", you're still left with a GeoDataFrame, albeit one with only one row. The distance operation is therefore performed elementwise, where first df and nigeria are aligned on their indices, and then the elementwise distance is taken between rows in the two dataframes. Since there are no common indices between the two, the result is a column of NaNs.

This works the same way as in pandas - if you add two series with different indexes together, even if one of them is length 1, you would get NaN as a result.

On the other hand, if nigeria is a shapely geometry object rather than a GeoSeries or GeoDataFrame, there is no index to align on, and geopandas will compare each element in the dataframe to the geometry. The pandas analogue here is adding a constant to a Series, which results in a valid sum of each element with the constant.

So, you can change your code to select out the shapely geometry object, then compare it to each element in df:

# use `.at` to extract the element instead of returning a length-1
# DataFrame. The result will be a shapely geometry.
nigeria = world.at[world['name'] == 'Nigeria', 'geometry']

type(nigeria)  # shapely.geometry.MultiPolygon

# compare each element in df to Nigeria's boundary
df['distance1'] = df.distance(nigeria.boundary)