def _calc_distances(df_prop, df_hosp):
"""
calculates the distance from each property to each hospital
Parameters
----------
df_prop: pandas.DataFrame
df_hosp: pandas.DataFrame
Returns
-------
pandas.DataFrame
"""
pos_prop = df_prop[['LATITUDE', 'LONGITUDE']].values
pos_hosp = df_hosp[['LATITUDE', 'LONGITUDE']].values
(prop_rows, prop_columns) = pos_prop.shape
(hosp_rows, hosp_columns) = pos_hosp.shape
distances = np.zeros((hosp_rows, prop_rows))
for k in range(hosp_rows):
distances[k, :] = distance_estimators.spherical_distance(pos_prop.copy(), pos_hosp[k].copy())
min_indices = np.argmin(distances, 0)
min_distances = np.amin(distances, 0)
hosp_names = df_hosp['name'].values
df_prop['NEAREST_HOSPITAL'] = hosp_names[min_indices]
df_prop['MIN_DISTANCE'] = min_distances
df_prop = df_prop[['LONGITUDE', 'LATITUDE', 'NEAREST_HOSPITAL', 'MIN_DISTANCE', 'AV_TOTAL',
'LIVING_AREA']].copy()
return df_prop
def _calc_distances(df_prop, df_school):
"""
calculates the distance from each property to each school
Parameters
----------
df_prop: pandas.DataFrame
df_school: pandas.DataFrame
Returns
-------
pandas.DataFrame
"""
pos_prop = df_prop[['LATITUDE', 'LONGITUDE']].values
pos_school = df_school[['LATITUDE', 'LONGITUDE']].values
(prop_rows, prop_columns) = pos_prop.shape
(school_rows, school_columns) = pos_school.shape
distances = np.zeros((school_rows, prop_rows))
for k in range(school_rows):
distances[k, :] = distance_estimators.spherical_distance(pos_prop.copy(), pos_school[k].copy())
min_indices = np.argmin(distances, 0)
min_distances = np.amin(distances, 0)
school_names = df_school['sch_name'].values
df_prop['NEAREST_SCHOOL'] = school_names[min_indices]
df_prop['MIN_DISTANCE'] = min_distances
df_prop = df_prop[['LONGITUDE', 'LATITUDE', 'NEAREST_SCHOOL', 'MIN_DISTANCE', 'AV_TOTAL',
'LIVING_AREA']].copy()
return df_prop
def sse_group(df, columns):
"""
Calculates the SSE for a cluster given a dataframe with a list of X columns
Parameters
----------
df: pandas.DataFrame
columns: list
Returns
-------
np.float
"""
distances = distance_estimators.spherical_distance(df[columns].values, np.array([df[columns].mean().values]))
distances *= distances
sse = sum(distances)
return sse
def sse_group(df, columns):
"""
Calculates the SSE for a cluster given a dataframe with a list of X columns
Parameters
----------
df: pandas.DataFrame
columns: list
Returns
-------
np.float
"""
distances = distance_estimators.spherical_distance(
df[columns].values, np.array([df[columns].mean().values]))
distances *= distances
sse = sum(distances)
return sse
def _calc_distances(df_prop, df_school):
"""
calculates the distance from each property to each school
Parameters
----------
df_prop: pandas.DataFrame
df_school: pandas.DataFrame
Returns
-------
pandas.DataFrame
"""
pos_prop = df_prop[['LATITUDE', 'LONGITUDE']].values
pos_school = df_school[['LATITUDE', 'LONGITUDE']].values
(prop_rows, prop_columns) = pos_prop.shape
(school_rows, school_columns) = pos_school.shape
distances = np.zeros((school_rows, prop_rows))
for k in range(school_rows):
distances[k, :] = distance_estimators.spherical_distance(
pos_prop.copy(), pos_school[k].copy())
min_indices = np.argmin(distances, 0)
min_distances = np.amin(distances, 0)
school_names = df_school['sch_name'].values
df_prop['NEAREST_SCHOOL'] = school_names[min_indices]
df_prop['MIN_DISTANCE'] = min_distances
df_prop = df_prop[[
'LONGITUDE', 'LATITUDE', 'NEAREST_SCHOOL', 'MIN_DISTANCE',
'AV_TOTAL', 'LIVING_AREA'
]].copy()
return df_prop
