Overview and use of Correlation Maps

We recently introduced a new method to display network mode maps in VizMAP and this blog focuses on discussing details of this map.

First, a recap on network mode maps.  Let us look at a very simple table of 5 companies vs IPC Main classes on a random sample of patents an applications on antivirus. The data in the tabular format is shown below.

If we represent that in network mode of VizMAP it appears as:

Key IPCs in the map are automatically placed in the center and company portfolios are organized around the Assignee-IPC relationship. The patents (smaller dots) are colored by Assignee and shading demarcates individual company portfolios.

Clicking on HO4L IPC Node will highlight all records that fall under this category:

As you can see few records of each Company are classified under the H04L. Now we could also shade the portfolios by IPC but that would overlap with the existing companywise shading and make the map complicated. So if you wanted to quickly see a similar correlation between all IPC codes and all Assignees without having to click on each node? That’s where the correlation map comes in. The correlation map for the same data is shown below: 

The map clearly shows how many records relate two nodes and the thickness of the line is also proportional to the number of correlating records. In addition to the data in the matrix the display also shows correlation between different IPC nodes. You can restrict the map to only the type of correlations you want to see and in the process bring out a visual that perhaps best represents the matrix data.

The correlation dosent need to be just Number of common records. We can have a co-citation map in the same display where the correlation is based on Number of common citing records. For example in the Fuel Additives patent set that we created for one of our Technology Insight Reports on Fuel Additives , a co-citation map for the Top-10 companies is shown below.

The map clearly shows that research on Fuel Additives happening at Chevron Corp and BASF AG is strongly related.