publicdata/j40-cejst-2
1
0
Fork 0
mirror of https://github.com/DOI-DO/j40-cejst-2.git synced 2025-02-23 01:54:18 -08:00
Code Issues Projects Releases Packages Wiki Activity Actions

Fix errors and improve code quality and readability in Health Scores (#1147)

Browse source 
* run black on health_score.py

* to_numpy() versus values - see https://pandas.pydata.org/pandas-docs/version/0.24.0rc1/api/generated/pandas.Series.to_numpy.html
This commit is contained in:
Saran Ahluwalia 2022-01-14 13:11:47 -05:00 • committed by GitHub
parent 667678f20e
commit 2604b66cf7
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 66 additions and 36 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

102
data/data-pipeline/data_pipeline/ipython/experiment_4_weighting/health_scores.py
View file

@ -8,10 +8,11 @@ import numpy as np
import pandas as pd import pandas as pd
from sklearn.preprocessing import MinMaxScaler from sklearn.preprocessing import MinMaxScaler
class HealthScores():
class HealthScores:
""" """
Calculates health scores by calling the final_scaled_data() method Calculates health scores by calling the final_scaled_data() method
Arguments: Arguments:
weights_1(np array) : weights of model 1. weights_1(np array) : weights of model 1.
weights_2(np array) : weights of model 2. weights_2(np array) : weights of model 2.
@ -21,9 +22,18 @@ class HealthScores():
is_weighted_average(boolean) : weights calculation methodology(default is True) is_weighted_average(boolean) : weights calculation methodology(default is True)
weightage(list) : weightage for each y-variable weightage(list) : weightage for each y-variable
""" """
def __init__(self,weights_1,weights_2,weights_3,multiplied_data,geoid,is_weighted_average = True,weightage = [0.25,0.25,0.5]): def __init__(
self,
weights_1,
weights_2,
weights_3,
multiplied_data,
geoid,
is_weighted_average=True,
weightage=[0.25, 0.25, 0.5],
):
self.weights_1 = weights_1 self.weights_1 = weights_1
self.weights_2 = weights_2 self.weights_2 = weights_2
self.weights_3 = weights_3 self.weights_3 = weights_3
@ -31,74 +41,94 @@ class HealthScores():
self.geoid = geoid self.geoid = geoid
self.is_weighted_average = is_weighted_average self.is_weighted_average = is_weighted_average
self.weightage = weightage self.weightage = weightage
self.weights = np.zeros(self.weights_1.shape).reshape(-1,1) self.weights = np.zeros(self.weights_1.shape).reshape(-1, 1)
def _get_weights(self): def _get_weights(self):
''' """
Averages weights of all the models and performs transformation so that sum of all weights will be equal to 1. Averages weights of all the models and performs transformation so that sum of all weights will be equal to 1.
Arguments : None Arguments : None
Returns : Averaged weights which totals to 1. Returns : Averaged weights which totals to 1.
''' """
weights = np.hstack((self.weights_1.reshape(-1,1),self.weights_2.reshape(-1,1),self.weights_3.reshape(-1,1))) weights = np.hstack(
weights = np.sum(weights,axis = 1) / weights.shape[1] (
self.weights_1.reshape(-1, 1),
self.weights_2.reshape(-1, 1),
self.weights_3.reshape(-1, 1),
)
)
weights = np.sum(weights, axis=1) / weights.shape[1]
self.weights = weights / np.sum(weights) self.weights = weights / np.sum(weights)
return self.weights return self.weights
def _weighted_average(self): def _weighted_average(self):
''' """
weighted average: weighted average:
Averages weights of all the models with specified weightage for each model. And performs transformation so that sum of Averages weights of all the models with specified weightage for each model. And performs transformation so that sum of
all weights will be equal to 1. all weights will be equal to 1.
Arguments : None Arguments : None
Returns : Averaged weights which totals to 1.(np array) Returns : Averaged weights which totals to 1.(np array)
''' """
weights = np.hstack((self.weights_1.reshape(-1,1) * self.weightage[0],self.weights_2.reshape(-1,1) * self.weightage[1],self.weights_3.reshape(-1,1) * self.weightage[2])) weights = np.hstack(
weights = np.sum(weights,axis = 1) (
self.weights_1.reshape(-1, 1) * self.weightage[0],
self.weights_2.reshape(-1, 1) * self.weightage[1],
self.weights_3.reshape(-1, 1) * self.weightage[2],
)
)
weights = np.sum(weights, axis=1)
self.weights = weights / np.sum(weights) self.weights = weights / np.sum(weights)
return self.weights return self.weights
def _health_score(self): def _health_score(self):
''' """
Converts data in (0 to 100)scale using min max scaler and multiiplying with 100. Converts data in (0 to 100)scale using min max scaler and multiiplying with 100.
Then it calculates health scores by multiplying with the weights Then it calculates health scores by multiplying with the weights
Returns : data frame with health score and x variables in (0 - 100)scale. Returns : data frame with health score and x variables in (0 - 100)scale.
''' """
columns = list(self.multiplied_data.columns) + ['health_scores'] columns = list(self.multiplied_data.columns) + ["health_scores"]
scaled_data = MinMaxScaler().fit_transform(self.multiplied_data.values) * 100 scaled_data = (
health_scores = np.dot(scaled_data,self.weights.reshape(-1,1)) MinMaxScaler().fit_transform(self.multiplied_data.to_numpy()) * 100
)
health_scores = np.dot(scaled_data, self.weights.reshape(-1, 1))
health_scores = MinMaxScaler().fit_transform(health_scores) * 100 health_scores = MinMaxScaler().fit_transform(health_scores) * 100
scaled_data = np.hstack((scaled_data,health_scores)) scaled_data = np.hstack((scaled_data, health_scores))
scaled_data = pd.DataFrame(data = scaled_data, columns = columns) scaled_data = pd.DataFrame(data=scaled_data, columns=columns)
return scaled_data return scaled_data
def final_scaled_data(self): def final_scaled_data(self):
""" """
Calls appropriate methods in class based on arguments. Concatenates geoids and health scores. Calls appropriate methods in class based on arguments. Concatenates geoids and health scores.
Arguments : None Arguments : None
Returns : Returns :
final_data(df) : dataframe with all health scores and geoids final_data(df) : dataframe with all health scores and geoids
weights_tables(df) : dataframe with each y-variable weights and averaged weights weights_tables(df) : dataframe with each y-variable weights and averaged weights
""" """
# final scaled data # final scaled data
if self.is_weighted_average: if self.is_weighted_average:
self._weighted_average() self._weighted_average()
else: else:
self._get_weights() self._get_weights()
scaled_data_100 = self._health_score() scaled_data_100 = self._health_score()
final_data = pd.concat([self.geoid,scaled_data_100],axis = 1) final_data = pd.concat([self.geoid, scaled_data_100], axis=1)
weights_table = pd.DataFrame({'phy_health_weights' : self.weights_1, 'mntl_health_weights' : self.weights_2, 'life_expectancy_weights' : self.weights_3, 'averaged_weights' : self.weights},index = self.multiplied_data.columns)
return final_data,weights_table
weights_table = pd.DataFrame(
{
"phy_health_weights": self.weights_1,
"mntl_health_weights": self.weights_2,
"life_expectancy_weights": self.weights_3,
"averaged_weights": self.weights,
},
index=self.multiplied_data.columns,
)
return final_data, weights_table