Introduction to Analysis
Analysis provides tools for working with data in the case database, such as finding how often a term appears in documents, or locating information about a specific person, place, or time. You can view key concepts in clusters of related documents, organize case data in a spreadsheet-like summary of rows and columns, create sets of documents for predictive coding, and assess small sets of documents to compare to larger sets.
In the Analysis section, you can create and manage cubes, mines, predictive models, populations, and samples. You can also stop concepts.
The Analysis section also includes the following dashboards: a Search Terms dashboard, which you can use to view statistics for search term hits in documents, and a Review Dashboard, which you can use to track the progress of an active review.
Cubes
Cubes are selected document data presented in a spreadsheet format in which the case data appears in rows and columns. Administrators perform tasks such as creating and managing cubes.
You can use a cube to identify documents that meet particular combinations of criteria, for example, custodians and date range. Review leads can use cubes to identify important documents quickly for early case assessment. When searching and reviewing in the application, you can view predefined cubes and copy documents from a cube to the Working list for further action.
Mines
A mine is a visual representation of important concepts in documents, presented as clusters of related documents. With mines you can explore groups of related documents and examine the relationships between the documents that are not obvious with a standard search or during an initial review. For example, you may realize that the same author created many documents in a case, but you may not know that only a few documents by this author are of interest to the case.
You can also use mines to create assignments in a case and route them to a review phase.
Predictive Coding
Predictive coding allows you to apply a positive or negative code to a large population of documents using reviewers' marks on a relatively small set of training documents. The predictive coding process uses a predictive model, which is trained to mimic the markings of expert human reviewers.
Populations and Samples
Populations and samples is an administrative feature that allows you to create a group of documents, called a population, and one or more subsets of that group, called samples. A sample consists of documents that the application randomly selects from the population using a mathematical algorithm. You can use samples to make inferences about the population, which helps with assessing a new case or monitoring the quality of coding. For example, if you observe that 25% of a sample is coded as privileged, then you could estimate that 25% of the population would be coded as privileged.
Review Dashboard
The Review Dashboard provides a progress bar, graph, and reviewer table that illustrate the progress of an active review. The information presented in the Review Dashboard is pre-populated to show all documents in the case. You can select a different document set, coding field, or team to update the information in the dashboard accordingly.
Search Terms
Use this feature to view statistics for search term hits in documents, and to view how each search term contributes to the overall population of hits for the family, along with other metrics. You can download the report to a spreadsheet (.csv file). Administrators can make this feature available to group leaders and group members.
Stop Concepts
A listing of concepts in a case. On this page, administrators can prevent concepts from appearing in mines, cubes, and the Map.