# Feature Engineering of mixed data type column

I have a data set in which I have to predict the price of a building. Among many features there is a feature called Availability which has two type values like :

"Ready To Move" and a date (e.g. "20-Jan", "18-May", etc.)


For example:

Availability           Price
------------           -----
Ready To Move           55
18-Apr                  38


How would you engineer this column to find its relationship with price and to draw other conclusions.

## 2 Answers

You have several choices. Among them try these:

PS: Please note that Spearman or Kendall Rank Correlations can also be applied to your categorical variables as they are ordered (dates).

You can create a new column with name "Number of days in which the building is going to be available" (Values will be (20-Jan / 18-Apr )-current date).