Datathon 2016
Marketplace Analytics
Marketplace Analytics and were growing Data-driven personalisation - - PowerPoint PPT Presentation
Datathon 2016 Marketplace Analytics and were growing Data-driven personalisation at Seek Matching Hello Antony Ugoni Head of Marketplace Analytics - SEEK SEEK identifies the need to become proactive in connecting candidates and
Marketplace Analytics
…and we’re growing
Antony Ugoni Head of Marketplace Analytics - SEEK
SEEK identifies the need to become proactive in connecting candidates and hirers.
Data-driven personalisation at Seek
Candidates to Job Matching – Weekly Roundup
You’ve applied for a job, here’s more like it. Here are some market insights for you too The results prove effectiveness and space for ‘matching’.
Target group Control group4x 6x
Antony Ugoni Head of Marketplace Analytics - SEEK
Datathon
Antony Ugoni Head of Marketplace Analytics - SEEK
Visualising the Data
Antony Ugoni Head of Marketplace Analytics - SEEK
Conceptual Schema
Antony Ugoni Head of Marketplace Analytics - SEEK
Job Searches
Column Name Comments
user_id If the user is logged in it is the user_id of the user executing the search, else it is the “cookied” user session id . search_id Unique ID representing a single search. raw_location Location text entered into the location field on website location_id When raw location is populated the GeoName ID is looked up against http://www.geonames.org/. This field is only populated if the looked up of the raw_location is found in the Geoname List latitude When raw location is populated the Latitude is looked up against http://www.geonames.org/ This field is only populated if the looked up of the raw_location is found in the Geoname List longitude When raw location is populated the Longitude is looked up against http://www.geonames.org/ This field is only populated if the looked up of the raw_location is found in the Geoname List query Search term entered in the query field on the website mobile_user True / False value representing whether the search was conducted on a mobile device. created_at UTC Date time the search event record was written to Elastic Search
Antony Ugoni Head of Marketplace Analytics - SEEK
Job Impressions
Column Name Comments
job_id The Unique ID of the Job to which the stats pertain. user_id If the user is logged in it is the user_id of the user executing the search, else it is the cookied user session id . session_id Cookied user session id. search_id The search_id from job_search from which this impression was created. search_ranking The ranked position of the Job in the search results page, starting from top to bottom. Numbering continues across pagination. mobile_user True / False value representing whether the search was conducted on a mobile device. created_at The UTC Date time when the job impression event record was written to Elastic Search
Antony Ugoni Head of Marketplace Analytics - SEEK
Job Clicks
Column Name Comments
job_id The unique ID of the Job which the user clicked on to view additional information. user_id If the user is logged in it is the user_id of the user executing the search, else it is the cookied user session id . session_id Cookied user session id. search_id The search_id from job_search if the job was clicked from a search results page created_at The UTC Date time the click event record was written to Elastic Search
Antony Ugoni Head of Marketplace Analytics - SEEK
Jobs Data
Column Name Comments
jobid The unique ID that represents a single job. Title The title of the job. raw_location The text description of the job’s location. location_id When raw location is populated the GeoName ID is looked up against http://www.geonames.org/ subclass The subclass id that the job is identified as belonging to salary_type The remuneration level that has been calculated for this job. All salaries have been annualised salary_min Minimum value of the Salary Range of the Job salary_max Maximum value of the Salary Range of the Job raw_job_type The job type of the job relating the amount of time the job requires Abstract A short description of the Job segment The market segment the job falls into. Values are 1 to 4 where 1 is low paid unskilled labour, 4 is high paid, high skilled careers.
Antony Ugoni Head of Marketplace Analytics - SEEK
Job Classification
Column Name Comments
class_id UniqueID that identifies a class by which a job can be classified. class_description Description of the class by which a job can be classified. sub_class_id UniqueID that identifies a sub_class by which a job can be sub classified. A job is classified against
Sub_class_description Description of the sub_class by which a job can be sub_classified
Antony Ugoni Head of Marketplace Analytics - SEEK
Contacts aalessi@seek.com.au wharrison@seek.com.au emazloumi@seek.com.au
Datathon 2016