Collaborative sensemaking system and method

Abstract

A system and method allowing machine analytics and multiple human analysts to work together in a collaborative fashion in the analysis of large amounts of data. The collaborative analysis promotes “sensemaking.” Sensemaking is the continued revising and expansion of a developing “story” so that it becomes more comprehensible. As the emerging story is refined, it should successfully incorporate more and more of the observed data and be more defensible with respect to contrary arguments.

Having described my invention, I claim: 
     
         1 . A method for collaborative sensemaking in which a group of human users work with machine analytics, comprising:
 (a) defining a data corpus comprising a large body of data;   (b) applying machine analytic software running on a computer to said data corpus, wherein said machine analytic software selects a docset from said data corpus, said docset being a much smaller subset of said data corpus;   (c) wherein said machine analytic software selects said docset on the basis of said docset containing entities that are previously defined as being of interest;   (d) wherein said machine analytic software annotates said docset by highlighting said selected entities previously defined as being of interest within said docset in a first color;   (e) creating a display of highlighted entities of said docset, wherein each highlighted entity includes said highlighting in said first color;   (f) providing a group of human users;   (g) defining criteria for the appropriate selection of entities within said docset by said human users;   (h) providing a software-based system for each of said human users to review said docset and highlight entities within said docset according to said defined criteria, using a second color that is different from said first color;   (i) providing software that creates a report including a visual presentation of said entities, wherein said software automatically applies color graduation to said visual presentation of said entities, with said color graduation indicating an increasing level of consensus as to entities highlighted by said users in said group, but preserving said highlights of said machine analytic software;   (j) wherein said color graduation is not visible to each of said users as said user is highlighting said entities, and said software only makes said color graduation visible after a user has added said highlights and said software has generated said report; and   (k) for each highlighted entity contained within said report, providing a link that leads back to a depiction of a source file from which said entity was originally taken.   
     
     
         2 . The method for collaborative sensemaking as recited in  claim 1 , wherein said docset includes at least one file type selected from the group consisting of text files, still image files, video image files, and audio files. 
     
     
         3 . The method for collaborative sensemaking as recited in  claim 1 , wherein said color graduation comprises using a progressively darker shade of a color to indicate an increasing level of consensus. 
     
     
         4 . The method for collaborative sensemaking as recited in  claim 1 , wherein said link leading back to a depiction of said source file comprises:
 (a) assigning a unique uniform resource identifier or uniform resource locator to a fragment containing said highlighted entity, said unique uniform resource identifier or uniform resource locator identifying a larger data source from which said fragment is taken; and   (b) when a user actuates a link in said report, using said unique uniform resource identifier or uniform resource locator to pull up a depiction of said larger data source so that said user may view said highlighted entity in context.   
     
     
         5 . The method for collaborative sensemaking as recited in  claim 1 , further comprising providing a linking function for said users, whereby each of said users can link one highlighted entity to another highlighted entity. 
     
     
         6 . The method for collaborative sensemaking as recited in  claim 1 , further comprising after said software selects said docset, allowing a user to review and edit said highlighting selections made by said machine before said docset is provided to said group of users. 
     
     
         7 . The method for collaborative sensemaking as recited in  claim 1 , further comprising said machine analytic software creating RDF triples using said highlighted entities, each of said RDF triples being a machine-readable tags with three parts including a subject, an object, and a predicate that indicates a relationship between said subject and said object. 
     
     
         8 . The method for collaborative sensemaking as recited in  claim 7 , further comprising creating a node graph depicting said relationships between said highlighted entities. 
     
     
         9 . The method for collaborative sensemaking as recited in  claim 5 , further comprising allowing a user to move among a series of linked entities by activating a next function. 
     
     
         10 . The method for collaborative sensemaking as recited in  claim 1 , wherein said entities to be highlighted include a pictorial image, a graphical image, a portion of text, and a portion of video. 
     
     
         11 . A method for collaborative sensemaking in which a group of human users work with machine analytics, comprising:
 (a) defining a data corpus;   (b) creating a defined software process for extracting a subset of said data corpus, said defined software process configured to be run on a computer as an automated process;   (c) applying said defined software process running on a computer to said data corpus, wherein said defined software process selects a docset from said data corpus, said docset being a smaller subset of said data corpus;   (d) wherein said defined software process annotates said docset by highlighting said selected entities previously defined as being of interest within said docset in a first color;   (e) creating a display of highlighted entities of said docset, wherein each highlighted entity includes said highlighting in said first color;   (f) providing a group of human users;   (g) defining criteria for the appropriate selection of entities within said docset by said human users;   (h) providing a software-based system for each of said users to review said docset and highlight entities within said docset according to said defined criteria using a second color that is different from said first color;   (i) providing software that creates a report including a visual presentation of said entities wherein said software automatically applies color graduation to said visual presentation of said entities, with said color graduation indicating an increasing level of consensus as to entities highlighted by said users in said group, but preserving said highlights of said machine analytic software;   (j) wherein said color graduation is not visible to each of said users as said user is highlighting said entities, and said software only makes said color graduation visible after a user has added said highlights and said software has generated said report; and   (k) for each highlighted entity contained within said report, providing a link that leads back to a depiction of a source file from which said entity was originally taken.   
     
     
         12 . The method for collaborative sensemaking as recited in  claim 11 , wherein said docset includes at least one file type selected from the group consisting of text files, still image files, video image files, and audio files. 
     
     
         13 . The method for collaborative sensemaking as recited in  claim 11 , wherein said color graduation comprises using a progressively darker shade of a color to indicate an increasing level of consensus. 
     
     
         14 . The method for collaborative sensemaking as recited in  claim 11 , wherein said link leading back to a depiction of said source file comprises:
 (a) assigning a unique uniform resource identifier or uniform resource locator to a fragment containing said highlighted entity, said unique uniform resource identifier or uniform resource locator identifying a larger data source from which said fragment is taken; and   (b) when a user actuates a link in said report, using said unique uniform resource identifier or uniform resource locator to pull up a depiction of said larger data source so that said user may view said highlighted entity in context.   
     
     
         15 . The method for collaborative sensemaking as recited in  claim 11 , further comprising providing a linking function for said users, whereby each of said users can link one highlighted entity to another highlighted entity. 
     
     
         16 . The method for collaborative sensemaking as recited in  claim 11 ,
 further comprising after said software selects said docset, allowing a user to review and edit said highlighting selections made by said machine before said docset is provided to said group of users.   
     
     
         17 . The method for collaborative sensemaking as recited in  claim 11 , further comprising said machine analytic software creating RDF triples using said highlighted entities, each of said RDF triples being a machine-readable tags with three parts including a subject, an object, and a predicate that indicates a relationship between said subject and said object. 
     
     
         18 . The method for collaborative sensemaking as recited in  claim 17 , further comprising creating a node graph depicting said relationships between said highlighted entities. 
     
     
         19 . The method for collaborative sensemaking as recited in  claim 15 , further comprising allowing a user to move among a series of linked entities by activating a next function. 
     
     
         20 . The method for collaborative sensemaking as recited in  claim 11 , wherein said entities to be highlighted include a pictorial image, a graphical image, a portion of text, and a portion of video.