The goal is to group the data by columns type, then within the groups identified, use Min-Max standardization to process the column called score and assign to a new column called normalization_score
import pandas as pd
import numpy as np
from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler()
here is the df looks like
here is the first try:
df["normalization_score"] = df.groupby(["type"]).apply(scaler.fit_transform(df["score"]))
and raise ValueError
ValueError: Expected 2D array, got 1D array instead:
array=[0.28571429 0.01639344 0.42857143 0.14285714 0.57142857 0.71428571
0.16393443 0.08196721 0.06557377 0.03278689 0.1147541 0.09836066
0.13114754 0.18032787 0.27868852 0.21311475 0.36065574 0.14754098
0.32786885 0.04347826 0.04918033 0.08695652 0.40983607 0.85714286
0.42622951 0.3442623 0.29508197 0.2295082 0.24590164 0.67213115
0.59016393 0.19672131 0.13043478 0.45901639 0.44262295 0.17391304
0.50819672 0.54098361 0.47540984 0.2173913 0.52459016 0.26086957
0.49180328 0.63934426 0.37837838 0.60655738 0.72131148 0.26229508
0.40540541 0.57377049 0.08108108 1. 0.62295082 0.6557377
0.55737705 0.68852459 0.13513514 0.27027027 0.59459459 0.34782609
0.24324324 0.52173913 0.73770492 0.05405405 0.02702703 0.43478261
0.43243243 0.30434783 0.7704918 0.60869565 0.51351351 0.78688525
0.2972973 0.10810811 0.48648649 0.35135135 0.18918919 0.80327869
0.56756757 0.73913043 0.69565217 0.81967213 0.32432432 0.65217391
0.7826087 0.75409836 0.83606557 0.7027027 0.7027027 0.7826087
0.7826087 0.7027027 0.83606557 0.56521739 0.7826087 0.45945946
0.31147541 0.83606557 0.16216216 0.7826087 0.83606557 0.62162162
0.7027027 0.7027027 0.39344262 0.37704918 0.7027027 0.47826087
0.83606557 0.7826087 0.83606557 0.7027027 0.67567568 0.7027027
0.83606557 0.83606557 0.83606557 0.39130435 0.7027027 0.83606557
0.21621622 0.70491803 0.83606557 0.54054054 0.7027027 0.64864865
0.7027027 0.7027027 ].
Reshape your data either using array.reshape(-1, 1) if your data has a single feature or array.reshape(1, -1) if it contains a single sample.
then i try this code followed by instruction
df["normalization_score"] = df.groupby(["type"]).apply(scaler.fit_transform(np.array(df["score"]).reshape(-1, 1)))
and it shows this error
File ~/opt/anaconda3/envs/pythonProject/lib/python3.10/site-packages/pandas/core/groupby/groupby.py:1248, in GroupBy.apply(self, func, *args, **kwargs)
1241 @Appender(
1242 _apply_docs["template"].format(
1243 input="dataframe", examples=_apply_docs["dataframe_examples"]
1244 )
1245 )
1246 def apply(self, func, *args, **kwargs):
-> 1248 func = com.is_builtin_func(func)
1250 # this is needed so we don't try and wrap strings. If we could
1251 # resolve functions to their callable functions prior, this
1252 # wouldn't be needed
1253 if args or kwargs:
File ~/opt/anaconda3/envs/pythonProject/lib/python3.10/site-packages/pandas/core/common.py:582, in is_builtin_func(arg)
577 def is_builtin_func(arg):
578 """
579 if we define an builtin function for this argument, return it,
580 otherwise return the arg
581 """
--> 582 return _builtin_table.get(arg, arg)
TypeError: unhashable type: 'numpy.ndarray'