|
| Variable | Original version | Recorded version |
|
| date_first/last_seen | Both variables are in format “yyyy-mm-dd” | 6 new variables were created, namely year_start/end; month start/end; and day_start/end |
| rent_or_sell property_type | “rent” and “sell” 1234 maisonette multiple rooms room studio | Recoded to binary on−0 rent, 1 for sale recoded completely to numbers-1, 2, 3, 4, 5, 6, 7, 8 |
| lister_type neighbourhood | Contains many character values with the name of the neighbourhood | An id number was assigned for the different neighbourhoods |
| lister_type | Contains the levels “owner,” “invester,” “builder,” “blank,” agency (looks like).” The value “agency (looks like)” is a mistake made during data collection. It represents in reality either investors or builders | Since no strict condition for the recognition between investor and builder was found, the value “agency (looks like)” was randomly replaced to be either builder or investor. New variables with codes from 1 to 4 were created |
| build_type | Originally the variable contains year and building material | The variable was split in two new variables-year_built and type_built |
| specials | Text variable in the format [\word1\,”\word2\,”\word3\,”…] | Binary variables for each word indicating the existence or lack of this feature |
| floor | Originally in the format for example “5 to 10” | Split in two new variables floor_new and total_floors |
|
