Home Â» How to Normalize Columns in a Pandas DataFrame

# How to Normalize Columns in a Pandas DataFrame

Often you may want toÂ normalizeÂ the data values of one or more columns in a pandas DataFrame.

This tutorial explains two ways to do so:

1.Â Min-Max Normalization

• Objective: Converts each data value to a value between 0 and 1.
• Formula:Â New value = (value â€“ min) / (max â€“ min)

2. Mean Normalization

• Objective: Scales values such that the mean of all values is 0 and std. dev. is 1.Â
• Formula:Â New value = (value â€“ mean) / (standard deviation)

Letâ€™s check out an example of how to use each method on a pandas DataFrame.

### Example 1: Min-Max Normalization

Suppose we have the following pandas DataFrame:

```import pandas as pd

#create DataFrame
df = pd.DataFrame({'points': [25, 12, 15, 14, 19],
'assists': [5, 7, 7, 9, 12],
'rebounds': [11, 8, 10, 6, 6]})

#view DataFrame
print(df)

points  assists  rebounds
0      25        5        11
1      12        7         8
2      15        7        10
3      14        9         6
4      19       12         6
```

We can use the following code to apply a min-max normalization to each column in the DataFrame:

```(df-df.min())/(df.max()-df.min())

points	        assists	        rebounds
0	1.000000	0.000000	1.0
1	0.000000	0.285714	0.4
2	0.230769	0.285714	0.8
3	0.153846	0.571429	0.0
4	0.538462	1.000000	0.0
```

The max value in each column is now equal toÂ 1Â and the min value in each column is now equal toÂ 0, with all other values ranging between 0 and 1.

### Example 2: Mean Normalization

Once again suppose we have the following pandas DataFrame:

```import pandas as pd

#create DataFrame
df = pd.DataFrame({'points': [25, 12, 15, 14, 19],
'assists': [5, 7, 7, 9, 12],
'rebounds': [11, 8, 10, 6, 6]})

#view DataFrame
print(df)

points  assists  rebounds
0      25        5        11
1      12        7         8
2      15        7        10
3      14        9         6
4      19       12         6
```

We can use the following code to apply a mean normalization to each column in the DataFrame:

```(df-df.mean())/df.std()

points	        assists   	rebounds
0	1.554057	-1.133893	1.227881
1	-0.971286	-0.377964	-0.087706
2	-0.388514	-0.377964	0.789352
3	-0.582772	0.377964	-0.964764
4	0.388514	1.511858	-0.964764
```

The values in each column are now normalized such that the mean of the values in each column is 0 and the standard deviation of values in each column is 1.

If a particular data point has a normalized value greater than 0, itâ€™s an indication that the data point is greater than the mean of its column. Conversely, a normalized value less than 0 is an indication that the data point is less than the mean of its column.