Looking To Transform Continuous Variables Into Categorical

Sample Data: id val1 val2 val3 val4 val5 val6 val7 ///+8yr NaN 0.0 2.0 NaN 1 3 23 ///1vjh NaN NaN NaN NaN NaN 7 62 ///4wu 3

Solution 1:

IIUC you have two questions. The first question of replacing values larger than 5 with 'larger than 5' can be achieved with boolean indexing and the second question of grouping can be achieved with pd.cut()

DEMO:

d = pd.read_clipboard()

Part 1

Obtaining the values that does not satisfy the larger than 5 criteria,

rest = d.loc[:,'val1':'val6'][~(d.loc[:,'val1':'val6'] >5)]
rest


   val1  val2  val3  val4  val5  val6
0   NaN   0.0   2.0   NaN   1.0   3.0
1   NaN   NaN   NaN   NaN   NaN   NaN
2   3.0   NaN   NaN   NaN   NaN   NaN

Obtaining the larger than 5 values

larger_than_5=d.loc[:,'val1':'val6'][d.loc[:,'val1':'val6'] >5]
print(larger_than_5)

   val1  val2  val3  val4  val5  val6
0   NaN   NaN   NaN   NaN   NaN   NaN
1   NaN   NaN   NaN   NaN   NaN   7.0
2   NaN   NaN   6.0   NaN   7.0   8.0

Updating with your logic,

larger_than_5[larger_than_5.notnull()] ='Larger than 5'
print(larger_than_5)

   val1  val2           val3  val4           val5           val6
0   NaN   NaN            NaN   NaN            NaN            NaN
1   NaN   NaN            NaN   NaN            NaN  Larger than 5
2   NaN   NaN  Larger than 5   NaN  Larger than 5  Larger than 5

Updating rest with the logic,

rest.update(larger_than_5)
print(rest)

   val1  val2           val3  val4           val5           val6
0   NaN   0.0              2   NaN              1              3
1   NaN   NaN            NaN   NaN            NaN  Larger than 5
2   3.0   NaN  Larger than 5   NaN  Larger than 5  Larger than 5

Replacing values of the original df with updated values as per logic 1

d.loc[:,'val1':'val6'] = rest
print(d)

        id  val1  val2           val3  val4           val5           val6  \
0  ///+8yr   NaN   0.0              2   NaN              1              3   
1  ///1vjh   NaN   NaN            NaN   NaN            NaN  Larger than 5   
2   ///4wu   3.0   NaN  Larger than 5   NaN  Larger than 5  Larger than 5   

   val7  
0    23  
1    62  
2   180  

---

Part 2

Obtaining bins

bins = np.arange(0, d['val7'].max()+1, 30)
bins

array([  0,  30,  60,  90, 120, 150, 180], dtype=int64)

Creating a new series

val7_groups = pd.cut(d['val7'], bins)
val7_groups

0       (0, 30]
1      (60, 90]
2    (150, 180]

Adding that to the dataframe

d['val7_groups'] = val7_groups
print(d)

        id  val1  val2           val3  val4           val5           val6  \
0  ///+8yr   NaN   0.0              2   NaN              1              3   
1  ///1vjh   NaN   NaN            NaN   NaN            NaN  Larger than 5   
2   ///4wu   3.0   NaN  Larger than 5   NaN  Larger than 5  Larger than 5   

   val7 val7_groups  
0    23     (0, 30]  
1    62    (60, 90]  
2   180  (150, 180] 

you can also set group labels by passing values to the labels parameter in pd.cut()