Dial visualizations with Relativity analytics

Relativity analytics includes a new tool called dial visualization that can help you get a quick overview of how documents are coded in a workspace.

Begin by selecting the Cluster visualizations dashboard.

Right click on a cluster and select 'View Dial'.

The relative proportion of each cluster will be easier to discern on the dial. The top level clusters will be arranged along the inner ring.

Clicking on a widget on the dashboard linked to clusters will cause the dial to update. So in this example if the documents tagged hot in the pie chart are clicked, the dial will update to correspondingly shade those clusters which contain more documents that are coded hot.

Clicking on an individual cluster will drill down to greater detail for a cluster.

If you right click, you will have the option to switch to the nearby cluster view. Here, it's possible to pinpoint clusters which are not similarly coded to the cluster you have focused on.

It's possible to remove the condition placed on the visualization by clicking on the drop down menu at the upper left:

Relativity Analytics for Small Data Sets

Relativity recommends the use of analytics (email threading; clustering; concept searching; etc.) even for projects where a large amount of ESI is not collected.

Two scenarios show how analytics can help a team get a handle on a matter in just a few days.

A company wanted to conduct an investigation after the sudden departure of two executives, who were suspected of taking proprietary information. In just 5 days the following steps were performed:

1. Email threading brought down the number of email messages to be reviewed by 30%.

2. Communication analysis quickly showed which employees were communicating directly with each other. The larger nodes indicated a greater number of communications.

3. Clustering grouped together conceptually similar documents in the data set.

4. Active learning - TAR - narrowed down a set of documents considered the most likely to be responsive.

5. Relativity's Case Dynamics includes a Timeline Builder that was used in a report to the client.

In a different project, a team was asked to review a data set for which no metadata was been included, and which had a lot of duplicate files. The data was quickly culled down to what needed to be reviewed in just 5 days.

1. The Find near duplicates tool was used to remove a significant percentage of documents.

2. Clustering detected documents groups which were clearly of no interest.

3. The Find Similar Documents analytics tool was used to find dozens of documents which are similar to a handful known to be relevant.

4. Pivot charts showed groups of electronic files which were outliers in the overall set.

5. The documents set was narrowed down further to only the custodians whose data was at issue in the matter. Relativity recently decided to use the term 'entity' for custodian - for some reason.

Concepts to Search for in Relativity

When running a concept search in Relativity, you do not have to base the search on a document from the workspace, or even an excerpt from a key document. Relativity recommends also basing concept searches on court filings (such as the complaint from the case), or another document such as a news article which is not in the review set. See this post.

A Relativity concept search will help you get the document you're looking for:

Avoid using only one word which will return "broad, unreliable results".