I find there are some features that may be "NULL" that the description doesn't mention such as prop_location_score_2. This indeed makes the parse and the competition more difficult.
\nAre there other features that may be "NULL" ? Also, could someone explain what do these NULLs mean?
", "messages": [ { "id": "30837", "postDate": "09/13/2013 08:16:53", "content": "I find there are some features that may be "NULL" that the description doesn't mention such as prop_location_score_2. This indeed makes the parse and the competition more difficult.
\nAre there other features that may be "NULL" ? Also, could someone explain what do these NULLs mean?
", "rawMarkdown": "", "votes": null }, { "id": "30842", "postDate": "09/13/2013 09:53:29", "content": "
srch_id
9917530
site_id
9917530
visitor_location_country_id
9917530
visitor_hist_starrating
505297
visitor_hist_adr_usd
507612
prop_country_id
9917530
prop_id
9917530
prop_starrating
9917530
prop_review_score
9902900
prop_brand_bool
9917530
prop_location_score1
9917530
prop_location_score2
7739150
prop_log_historical_price
9917530
position
9917530
price_usd
9917530
promotion_flag
9917530
srch_destination_id
9917530
srch_length_of_stay
9917530
srch_booking_window
9917530
srch_adults_count
9917530
srch_children_count
9917530
srch_room_count
9917530
srch_saturday_night_bool
9917530
srch_query_affinity_score
635564
orig_destination_distance
6701069
random_bool
9917530
comp1_rate
235806
comp1_inv
254433
comp1_rate_percent_diff
184907
comp2_rate
4040633
comp2_inv
4251538
comp2_rate_percent_diff
1109847
comp3_rate
3059273
comp3_inv
3292221
comp3_rate_percent_diff
944007
comp4_rate
620099
comp4_inv
692471
comp4_rate_percent_diff
264213
comp5_rate
4444294
comp5_inv
4720833
comp5_rate_percent_diff
1681006
comp6_rate
482487
comp6_inv
524145
comp6_rate_percent_diff
193312
comp7_rate
631077
comp7_inv
713175
comp7_rate_percent_diff
277838
comp8_rate
3819043
comp8_inv
3960388
comp8_rate_percent_diff
1225707
click_bool
9917530
gross_bookings_usd
276592
booking_bool
9917530
\n
in above, the columns contain null value if the number is less than 9917530.
", "rawMarkdown": "", "votes": null }, { "id": "30849", "postDate": "09/13/2013 12:03:24", "content": "The actual semantic of a missing value is different for different columns:
\n- visitor_hist_starrating, visitor_hist_adr_usd: there was no previous purchase history associated with a given user,
\n- prop_review_score: missing data in our data warehouse,
\n- prop_location_score2: it was not possible to compute this score; in this case it is valid to replace missing values with a small score (even better: replace it with a minimum value taken over all the hotels in the same destination, i.e. the same srch_destination_id),
\n- srch_query_affinity_score: there was no search engine query associated with a given impression,
\n- orig_destination_distance: for some reasons we did not have information on a physical location of a user (i.e. missing IP address or not possible to associate IP address with a physical location),
\n- comp*_rate, comp*_inv: we did not have information respectively on rate and inventory from our competitors,
\n- comp*_rate_percent_diff: we did not have information on the rate from our competitors, or comp*_rate = 0,
\n- gross_bookings_usd: only a subset of hotels got booked.
\nAdam
\n\n
", "rawMarkdown": "", "votes": null } ], "comments": [ { "id": 30842, "author_name": "zhaogang", "author_url": "", "post_date": "09/13/2013 09:53:29", "content": "
srch_id
9917530
site_id
9917530
visitor_location_country_id
9917530
visitor_hist_starrating
505297
visitor_hist_adr_usd
507612
prop_country_id
9917530
prop_id
9917530
prop_starrating
9917530
prop_review_score
9902900
prop_brand_bool
9917530
prop_location_score1
9917530
prop_location_score2
7739150
prop_log_historical_price
9917530
position
9917530
price_usd
9917530
promotion_flag
9917530
srch_destination_id
9917530
srch_length_of_stay
9917530
srch_booking_window
9917530
srch_adults_count
9917530
srch_children_count
9917530
srch_room_count
9917530
srch_saturday_night_bool
9917530
srch_query_affinity_score
635564
orig_destination_distance
6701069
random_bool
9917530
comp1_rate
235806
comp1_inv
254433
comp1_rate_percent_diff
184907
comp2_rate
4040633
comp2_inv
4251538
comp2_rate_percent_diff
1109847
comp3_rate
3059273
comp3_inv
3292221
comp3_rate_percent_diff
944007
comp4_rate
620099
comp4_inv
692471
comp4_rate_percent_diff
264213
comp5_rate
4444294
comp5_inv
4720833
comp5_rate_percent_diff
1681006
comp6_rate
482487
comp6_inv
524145
comp6_rate_percent_diff
193312
comp7_rate
631077
comp7_inv
713175
comp7_rate_percent_diff
277838
comp8_rate
3819043
comp8_inv
3960388
comp8_rate_percent_diff
1225707
click_bool
9917530
gross_bookings_usd
276592
booking_bool
9917530
\n
in above, the columns contain null value if the number is less than 9917530.
", "votes": null, "replies": [] }, { "id": 30849, "author_name": "adamwoz", "author_url": "", "post_date": "09/13/2013 12:03:24", "content": "The actual semantic of a missing value is different for different columns:
\n- visitor_hist_starrating, visitor_hist_adr_usd: there was no previous purchase history associated with a given user,
\n- prop_review_score: missing data in our data warehouse,
\n- prop_location_score2: it was not possible to compute this score; in this case it is valid to replace missing values with a small score (even better: replace it with a minimum value taken over all the hotels in the same destination, i.e. the same srch_destination_id),
\n- srch_query_affinity_score: there was no search engine query associated with a given impression,
\n- orig_destination_distance: for some reasons we did not have information on a physical location of a user (i.e. missing IP address or not possible to associate IP address with a physical location),
\n- comp*_rate, comp*_inv: we did not have information respectively on rate and inventory from our competitors,
\n- comp*_rate_percent_diff: we did not have information on the rate from our competitors, or comp*_rate = 0,
\n- gross_bookings_usd: only a subset of hotels got booked.
\nAdam
\n\n
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "30837": "", "30842": "", "30849": "" }, "source": "meta" }