This data contains historical matching between prospective customers and agents. Each row represents an assignment of an agent to a buyer. It is described here with the outcome rows include. When being used for testing your solution, the final two columns are omitted.
When a customer comes in search of a real estate agent, they provide some limited information. First off, they provide geographic information for where they are from, and where they are interested in buying. In some cases, they also provide a minimum and maximum prices. A few other bits of information are also automatically logged. Thus customer is then paired with an agent, and it is your task to evaluate a number of pairings with unknown outcomes. Remember that the goal of your algorithm is to rank the matchings with good outcomes higher than those with bad outcomes.

Download Matchings.csv
Fields:
  1. MatchedLeadID -- A unique ID for the matchiing
  2. REAgentID -- the real estate agent matched
  3. LeadType -- always 'Buy' in this case
  4. LeadCompletedDate -- the date the lead was assigned to an agent
  5. InterestedZip -- the zip code a customer is interested in buying in
  6. InterestedCity -- the city a customer is interested in buying in
  7. InterestedState -- the state a customer is interested in buying in
  8. ConsumerZip -- the customer's current zip code
  9. ConsumerCity -- the customer's current city
  10. ConsumerState -- the customer's current state
  11. MinimumPrice -- if specified by the customer
  12. MaximumPrice -- if specified by the customer
  13. MarketingChannel -- this specifies broadly how the customer found out about this site (which finds agents for individuals)
  14. PartnerID -- a unique ID for the marketing partner that the customer came to the site because of
  15. MarketingPartner -- the name of the company or website for the PartnerID
  16. IsReplyXMLPost -- a method used by some marketing partners
  17. IncentiveFlag -- this flag indicates whether or not a lead has a rebate associated with it
  18. QFormType -- this specifies the type of a particular form that the customer fills out
  19. LeadStatus -- this gives the status of the lead. The good outcomes are a set of the statuses
  20. Success -- a 1 is a good outcome, a 0 is a bad outcome