When Watson™ Explorer Engine gathers reporting data, it inserts it into a large SQL database, in multiple tables. The tables that it inserts data into will vary according to which reporting options have been enabled in the Reporting tab of a project. Here is an overview of the tables that may exist in your reporting database.
Global values (one value per "session"):
- reporting - this table consists of the primary key and the timestamp of the event.
- application - a simple string specified at the project level to differentiate multiple sets of projects.
- query - the query entered by the user, normalized for spaces and capitalization.
- pageid - the value of the "page" or "id" cgi parameter.
- tb - toolbar usage tracking (records the "tb" CGI parameter).
- gut - (general user tracking) the value can be manipulated in the project's main macro.
- user - if the query was performed by a user logged into Watson Explorer Engine, this is the name of the user.
- project - the name of the project.
- ip - the user's IP address (value of REMOTE_ADDR CGI environment variable).
- remotehost - the value of the REMOTE_HOST CGI environment variable.
- referrer - the value of the HTTP_REFERER CGI environment variable.
- servername - the value of the SERVER_NAME CGI environment variable.
- ua - the full user-agent string.
- uaplatform - the parsed platform (operating system) from the user-agent, if recognized.
- uabrowser - the parsed browser name from the user-agent, if recognized.
- uabrowserfull - the parsed full browser name (name with platform) from the user-agent, if recognized.
- uabrowserversion - the parsed browser version number from the user-agent, if recognized.
- ualanguage - the language specified by the user-agent, if recognized.
- ualocale - the locale specified by the user-agent, if recognized.
Other global values:
- term - an individual term in the user's query (there may be multiple values per session).
- cluster - the descriptor of a cluster in the output (when track-clusters is enabled).
- spelling - the suggested spelling correction, if presented to the user.
- boost - the number of boosts for this query.
Per-source values (the "key" column for each entry is the source name):
- responsetime - the amount of time elapsed between the start of processing and the end of the last fetch (per source).
- fetchingtime - the amount of time elapsed between the start of the first fetch and the end of the last fetch to complete (per source).
- parsingtime - the total amount of time spent parsing (per source).
- numresults - the number of results returned (per source).
Per-result-click values (the "key" column for each entry is a unique identifier generated for that click):
- url - the url of the result that was clicked by the user.
- rank - the rank (from the source) of the result click by the user.
- paid - the value "1" if the clicked result was a paid listing.
- resultsource - the name of the source from which this result was fetched.
- resultlabel - the descriptor of the cluster containing the clicked result, if it is in a cluster that was expanded by the user.
- resultcluster - the rank of the cluster containing the clicked result, if it is in a cluster that was expanded by the user.
- clusterposition - the rank of the cluster that was clicked by the user.
- descriptor - the descriptor of the cluster that has been clicked by the user.