Pandas: reverse Liker survey answers

Question

I have the following problem:

I have a survey that contains a large number of answers to likert questions, like so:

id | Q1 | Q2 | Q3
1     5    3    1
2     3    4    1
3     2    3    1

The problem is that not all questions are asked in 'the same direction'. So a answer of '5' in Q1 would indicate a positive answer. But a 5 in Q2 would mean a strongly negative answer.

We are currently re-encoding all questions by hand (thus replacing all Q2 5's with 1's, etc) but I was wondering if there is a quicker way to solve this.

I thought about dividing all answers by 5 and then subtracting 1, but that never gives me whole numbers. Math isn't really my strongpoint here, so I was wondering if someone here could help me out

Answer 1

If I understood you correctly, you can do it so:

df['Q2'] = df['Q2'].map({1:5, 2:4, 3:3, 4:2, 5:1})

Input:

   Q1  Q2  Q3
0   5   3   1
1   3   4   1
2   2   5   1

Output:

   Q1  Q2  Q3
0   5   3   1
1   3   2   1
2   2   1   1

Answer 2

You can subtract 6 from column Q2, or use rsub what subtract from right side:

print (df)
   Q1  Q2  Q3
0   5   3   1
1   3   4   1
2   2   5   1

df.Q2 = 6 - df.Q2
#same as
#df.Q2 = df.Q2.rsub(6)

If performance is important subtract by numpy array:

df.Q2 = 6 - df.Q2.values

Or:

df.Q2 = df.eval(' 6 - Q2')

Or:

import numexpr

x = df.Q2.values
df.Q2 = numexpr.evaluate('(6 - x)')

---

---

print (df)
   Q1  Q2  Q3
0   5   3   1
1   3   2   1
2   2   1   1