Report:TotalPatent/Viewing Results/Analyzing Results

Analyzing Results

Analysis functions in TotalPatent may only be used on documents that have been saved into as result sets into the Analytics section, and results can be saved for analysis from either the results list or the Work Folder record list. Users can select the "Analyze" option from either the result list of a search or from the record list of a selected folder, and selecting this option will open the "Save to Analytics" window, where users can choose to:

• Go to the Analytics section immediately to visualize the results.
• Go to the Analytics section immediately to compare results.
• Save the result set to the Analytics section to be used later.

The "Save to Analytics" window.

Users must save the search result set or Work Folder set to the Analytics section before the set is selectable from the list of result sets that can be analyzed and compared through Analytics.

Although a maximum of 3,000 documents are visible when viewing search hits, up to 20,000 documents can be saved into a Work Folder. The analysis functions in TotalPatent may be performed on either a search result set of up to 3,000 or a Work Folder set of up to 20,000.

The analysis features are initiated by choosing the “Analyze” option above the results list or the Work Folder list. The user will then be prompted with the "Save to Analytics" option window (shown above), and the user can choose to save the set for later or analyze/compare the set now. Alternatively, the user can also go directly to the analytics section by choosing the "Analytics" tab. Under the Analytics tab, users can choose from two available options: "Visualize" or "Compare."

Visualize

After selecting the "Visualize" tab, this screen will display the list of result sets available for analysis and a variety of analysis options, as shown in the figure below.

The visualization options under the "Analytics" tab in TotalPatent.

The list of available result sets includes the set name, where the set was saved from, and when the set was created. Users can select one set at a time to create visualizations for it. After selecting a set, a loading menu will display the message "The Analytics data for this folder is loading. Please wait."

The loading page for the analytics feature.

When the set is finished loading, a menu will display all available analysis options for the set. This menu can be used to construct various charts from the source data in the work folder, and users can choose from the following options:

• Users may choose to analyze all data in the folder, or to continue analyzing sub-sets of data. Any data point on a graph, once selected by the user, is eligible for further drill-down analysis to be performed on that subset of data, including graphical analysis on that sub-set alone. Users may choose from the following list of fields, and (depending on the first field selected) users may also choose a second field to include in the analysis:
  

• Authority
• Kind Code
• Inventor Name
• Assignee Name
• Normalized Assignee
• Publication Date
• Filing Date
• Priority Date
• Granted Date
• US Class
• IPC Class
• ECLA Class
• Most Recent Legal Code
• Legal Status – Total Positives
• Legal Status – Total Negatives
• Attorney/Registered Agent
• US Examiner

• Users can choose to restrict the graph to a certain number of “top” results, or to include all the data points in the graph. The options available from this drop down menu are 3, 5, 10, 20, 50, 100, and ALL.

• Users can add a date limitation to filter the data to only the desired date ranges. Any date range may be specified, and pre-defined options such as “previous 10 years” and “previous 20 years” are included in the drop-down menu.

• Finally, users may define the chart type to be used to graph the data. Available charts will change based on the types of data selected for inclusion into the graph. Possible selections include simple line, bar, and column charts, stacked bar and column charts, and pie and “bubble” charts.

A few examples of analysis charts produced by TotalPatent are shown below, using a data set constructed by a search for all documents listing the IPC class H04L 27/38 defined as: Modulated-carrier systems: Carrier systems characterised by combinations of two or more of the types covered by groups; Amplitude- and phase-modulated carrier systems, e.g. quadrature-amplitude modulated carrier systems; Demodulator circuits; Receiver circuits.

The first chart shows the top 50 assignees that are represented in the data set, broken down by assignee (the y-axis). Within each assignee's bar, the results are also color coded by the country of publication. In the graphs below, the company name Matsushita Electric appears several times, due to variations of the entity name (this issue can be fixed with the Merge feature, discussed in further detail below).

Example of analysis chart: the top 50 results from the data set (a single IPC class) sorted by assignee, subdivided with colors by publishing authority.

Next is a pie chart showing the data set broken down into the top 20 countries of publication.

Example of analysis chart: A pie chart showing the contents of the IPC class, broken down by top 20 patenting authorities.

As a final example, the screenshot below shows a "bubble chart" displaying the data set broken down by assignee vs. patenting authority. The size of the bubbles shows the relative size of the groups in comparison to one another. The epicenter of the bubble is positioned at the point of the grid showing the x and y values. In the graph below, the largest visible bubble (in the lower left corner) corresponds to the number of JP documents with no listed assignee, because the center of the bubble lines up with those two values on the x and y axis.

Example of analysis chart: A bubble chart showing assignees vs. patenting authority.

As of March 2012, a new "Merge" tool was added to the Analysis platform, where users can select the "Combine Items" link at the top of the chart to "manually 'normalize' data fields such as Assignee Name, Inventor Names, Classes, etc."^[1]

Select the "Combine Items" link to merge particular data points in the chart.

After selecting the "Combine Items" link, users will be presented with a window where they can select from all available data points for each field. The user can select multiple data points representing the same item, such as multiple assignee names representing the same company (as in the screenshot below). The user then must enter a new name for the combined items (such as "Agilent" to represent the three merged assignee names in the screenshot below), and the user can also specify whether this combination should be applied automatically to all analysis charts for this data set from this point forward.

Select to merge all assignee names representing the same company.

After you select "ok", the "Merge" window will close, and the combination will automatically be applied to the analysis chart. The user can choose to undo the last combination or edit the combination settings. When a user mouses over a section of the chart that represents a set of merged data points, the label that appears will display the newly assigned name for all merged data points, with the individual data point names listed below (see the screenshot below).

Multiple assignees are merged under a single name.

Once the analysis charts have been created, they are interactive: a user may select a particular data point of interest, which will begin to flash. A list of all individual patent documents contained within that data point will then appear below the chart, where they can be viewed in more detail. Users can select to "Retrieve Documents" and file all selected documents relating to that section of graph in a new work folder. After clicking on a portion of the data within an analysis chart, that data can also be analyzed individually by choosing the “analyze selected” button at the top of the analysis menu. Any further analysis actions will be performed on that data subset only, providing a “drill-down” feature. The figure below shows an example of the system displaying individual records associated with a section of a graph.

A list of documents that make up a subset of the data displayed on an analysis chart.

Finally, users can select the "Print Chart" or "Save Chart" options to print a version of the chart or save the chart in PNG format.

A printed version of a chart.

Compare

Users also can select the "Compare" tab (or select the comparison option under the "Save to Analytics" window from the results list or Work Folder) to bring up a list of all search results, work folders, and document retrieval results that have been saved to the Analytics section. From this list, users can select up to three sets and then choose the "Compare Selected" option to generate a Venn diagram visualization that illustrates the total amount of documents per set, all documents intersecting between two or more sets, and documents unique to particular sets. Users can print or save this Venn Diagram by selecting the "Print Chart"/"Save Chart" options.

A Venn diagram comparing three different result sets.

Select either the visualized portion of the set within the Venn diagram or the labels that describe how many documents are within each section of the Venn diagram, and the documents associated with the section of the Venn diagram will appear in a list below the comparison chart. The list below the comparison chart resembles a normal hit list, with:

• Various sorting and field display options
• Options to view the documents in list format or one document at a time
• Check boxes to select a document individually or select all documents
• Export and save options: email, print, download, or save selected results to a Work Folder
• Visualize the set using the analysis options under the "Visualize" tab

A list of results corresponding to a section of the Venn Diagram, with the option to "Visualize this set" with the analysis tools under the "Visualize" tab.

Editor's Note:

TotalPatent has taken some steps to improve the data cleaning options integrated within the analysis section. The Merge option is a particularly useful feature for normalizing data points within a field where multiple variations of a single name will result in multiple data points for one entity (such as assignee name). It would be helpful if TotalPatent included the additional option to delete data points altogether, since there is no simple way to “eliminate” a skewed data point while retaining all the rest of the data. For example, in the bubble chart screenshot above, the number of JP documents with a blank assignee field at 493 overwhelms most of the other data points on the chart. To get rid of this bubble while keeping all the others, the user would have to go back to the work file and delete the 493 offending patents.

Alternatively, the user could also complete the following steps to view the result set without the offending data point:

• Select the offending data point in the chart to view a list of all corresponding documents.
• Retrieve the documents associated with that data point into a separate work folder.
• Save that work folder to the analytics platform.
• Under the comparison tab, select to compare the original data set and the new data set (comprised of the documents associated with the offending data point).
• Within the Venn diagram, select to view all documents unique only to the original data set.
• Choose the visualization option for the new data set.

Although the technique described above is less time-consuming than deleting 493 patents from a set, the technique still requires multiple steps and is time-consuming. It would be much easier for the user to have the simple, direct option of deleting selected data points and the corresponding documents entirely.

The above technique utilizing the "Compare" tab also illustrates how the "Compare" option can be used as a method for combining sets using Boolean Operators AND/NOT, but the method still requires multiple confusing steps. Users of TotalPatent would benefit from the simple option to combine sets using Boolean operators.

One other minor obstacle to the analytics feature in TotalPatent is that every time it is accessed, the user must wait for the data from the results set to load into the TotalPatent analytics program. The wait time is usually only around 10-15 seconds, but it happens anytime a user navigates away from the analytics feature, and then wishes to return.

The analytics features are provided more for at-a-glance summary analysis than in-depth analysis projects and users are likely to export patent data for use in another graphing program (such as Microsoft Excel) rather than use this tool for large and in-depth analysis projects so the wait time is probably not an obstacle to most patent researchers.