BasicsLogin & Analysis Types
Understand the Radar
Patents in the data set are text mined and grouped together in clusters based on semantic similarities, or individual plots are created if an unclustered analysis is selected. Then they are run through multi-dimensional scaling and precisely positioned on a 2D radar based on their differences. Each dot represents one cluster or plot, with the size of the cluster reflecting the number of patents it contains. The closer the clusters the more similar their contents are. The XY axes have no particular meaning: distance is the focus of the data landscape.
Contour lines are created based on document density in areas of the radar, indicating possible semantic similarities or other levels of connectivity.
The analysis is automatically based on the company or organization that has the most patents in the data set: the Target Company. The others on the target list are the ones with the closest centers of gravity (overall document similarity) and filing ranges to the target company.
Create an Analysis
Select Gravity & Trend Analysis → Patent Office → Status/Language → Patent limit → Use Vision...
Japan = JPO
United States = USPTO
Europe = EPO
WIPO = WIPO
Multiple = JPO Granted and Application, or USPTO and EP Granted and Application plus WIPO Application
Application or Granted.
For WIPO, only Application is available
Create a Data Set to Analyze
Name: enter the name you'd like to give the data set.
Search type: select Text (concept search) or Patent Number.
Search query: enter up to 520K characters in the query box for a concept search, or enter patent numbers for the basis of the analysis. For best accuracy with text searches, the more words the better. If less than 5 are entered, any document containing any of the words can be included in the results (creating noise). Note that the tool mainly extracts and utilizes nouns, verbs, and adjectives for the analysis.
noteFor WIPO and EPO patent searches, the type of document must be included in the number, e.g. WO1991011363A1
If patent numbers are entered in the query box, the text in the patent(s) will be mined and similar documents will be included in the results. To analyze a list of specific patents only, use the Upload Patent List File function (see below)
- Part of Document to Search:
- Full Text searches the entire document.
- Abstract and Claims searches just that part of the document.
- Full Text with emphasis on Abstract and Claims searches the entire document and puts extra weight (priority) on the Abstract and Claims.
- Number of patents: enter the maximum number of patents you'd like to include in the data set, up to 100K.
- Filter by Bibliographic Data:
- NOTE: Using quotation marks when filtering helps reduce noise in the data set, e.g. "fuel cell" or "Apple, Inc". You can also use search operators AND, OR, (-) negation, ( ), and "string literal".
- Title: use to search by words in a patent title.
- Applicant: applicant/assignee refers to the company or organization name, e.g. Hitachi or Harvard University.
- Latest Assignee: refers to the most recent owner of a patent. Only available on US-A, US-R, JP-A, and JP-R.
- Inventor: full names must be input with last name, then a comma, then first name. e.g. EDISON, Thomas.
- IPC: single IPCs must be entered as, e.g., B64C. If adding IPC classification group/subgroup, it must be in quotation marks with spacing between the code and number. Use two spaces if the group/subgroup has 4 digits, e.g., “B64C␣␣15/12”, and three spaces if the group/subgroup has 3 numbers, e.g. "B64C␣␣␣9/00". If doing multiple, follow the same spacing rules with AND or OR in between, e.g., “B64C 15/12” AND “B64C 9/00”.
- Use only primary IPC: extracts only patents with the entered IPC if it is the primary (main) IPC.
Time range: enter the year range for the search.
Or Upload Patent List File: upload a .xlsx, .csv, .tsv, or .txt file of the patent numbers you'd like to analyze.
noteOr Upload Patent List File: The list must contain patent numbers corresponding to the status (Application or Granted) and issuing authority (e.g. USPTO-A, EPO-G, etc.) of the selected database. [Download a sample] (US-R)
If the uploaded list contains fewer patents than the analysis maximum, the results will contain the patents on the list and others pulled from the database according to text similarities. To create an analysis with only the patents on the uploaded list, enter the total number of patents on the list in the Sampling Threshold field (Data Sets → Analyze → Sampling Threshold)
Delete and change package: to stop this process and return to the home page, click [X] next to the current package. All data entered for this search will be erased.
Create Data Set: click to create the data set.
Data Sets: view all of your data sets here. To analyze, select one or multiple data sets you want to include in an analysis.
Under Analyze, confirm the Data settings:
Analysis Name: enter the name you'd like to give your analysis. If left blank, an analysis number will automatically be assigned.
Data Set: selected data sets will be listed here. Total Documents gives the cumulative number of patents in all selected data sets.
Part of Document to Analyze: confirm the part of the patent to be used for the analysis.
Sampling Threshold: confirm the number of patents to include in the analysis.
noteIf the total number of patents in the selected data sets is higher than the maximum allowable on your account, a calculated sampling will be extracted for the analysis
Use Options to apply custom rules to the analysis calculation.
Name Normalization is the aggregation of IPC, company, or inventor names. To apply a new rule, create and upload a file under the More dropdown → Name Normalization. Sample File
Cluster Strength increases or decreases the cluster volume by adjusting the similarity threshold.
Strong → fewer clusters with more documents per cluster.
Weak → more clusters with fewer documents per cluster.
Cluster Count allows you to specify the exact number of clusters created during the calculation process. To create an unclustered analysis, enter the Total Documents number of the data sets you want to analyze. See Data → Data Set → Total Documents.
Word Importance is the adjustment of the level of relevance of certain words during the similarity calculation and clustering process. Words can be emphasized, weakened, or excluded. To apply a new rule, create and upload a file under the More dropdown → Word Importance. Sample File
Word Grouping puts synonymous words together and runs them as one in the calculation process. To apply a new rule, create and upload a file under the More dropdown → Word Grouping. Sample File
Indexical Analyzer controls the parameters of the 2D visualization according to the number of documents in the data set and the scope of their contents. If a data set broadly covers a topic or field, e.g. renewable energy, set it to broad. If it is specific, e.g. solar power energy storage and transfer, set it to specific. For patent volume, low is less than 1,000. Average is 1,000-5,000. High is 5,000+.
Advanced Options: adjust calculation settings for Contour Lines, Gravity Distance Transition, and Noteworthy Areas.
Click Run Analysis to start.
Your Analysis is in Progress: view the status of your analysis here. When complete, you will be notified by email. You can leave this page without affecting the progress. You can check the status under My Analyses.
noteDepending on the document volume, it can take a few minutes to several hours to run the analysis
After running an analysis, it will be listed on the Create/Analyses page under My Analyses. You can hide, download, share, and delete analyses on this list.
Under My Analyses, select the analysis you'd like to view.
1. Return to your list of analyses or create a new one
2. Manage your data sets
3. Manage your word importance, word grouping, and name normalization files, and view your browsing history and license usage
4. Switch to Trend Analysis
5. Create a data subset of the analysis
7. Share the analysis: to enable, click Share, and then send a link of the analysis URL to others who have their own account
8. Areas: click to activate the Create Areas function
9. Areas: click to toggle Select Areas function
11. Cluster size control
12. Full screen
13. Manage the Target Company list, and access advanced visualization settings
14. Clear all: work on the radar is automatically saved. Click to return the radar to its original state
Find on Radar
Analysis Information: view the metadata of the analysis and access contained data sets.
Analysis Statistics: view statistical data of the analysis.
Layers: control the visible layers on the radar.
Keyword Density Areas: highlight major (high density), growing, or sparse (low density surrounded by high density) areas on the radar.
noteKeyword Density Area labels are off by default. To activate them, go to Layers → Keyword Density Areas and select.
M: Major, G: Growing, S: Sparse, N: Newer, U: Unique, W: Weakness, P: Strength
Cluster Trend: highlight the clusters containing the newest patents (based on issue date), largest number of patents, and closest to the centroid of the analysis. Click to highlight on the radar. Single click on the circle next to the cluster number to change the highlight color.
SWOT: view the target company’s strong and weak areas (based on document density and filing trend compared to others in the data set), its nearest assignee (the one with the most overall similar documents) and its caution assignee (the one moving fastest towards its center of gravity). Click to highlight on the radar. Single click on the circle next to the cluster number to change the highlight color.
Cluster Select: lists the cluster(s) you have clicked on the radar, including key words, years, and patents.
Cluster Select History: lists the history and access count of clusters you have clicked on the radar.
- Density: shows the integration density of each cluster of the target company. X and Y axes are the integration density of the target compared to others in the data set. The blue line is a least-squares method regression line. The red dotted lines are 3-sigma lines. The plots above the blue line are the target's strengths, those below are its weaknesses. The further from the axes origin the more unique. (See SWOT above for more details)
- Gravity Distance Transition: shows the changes in distance between the target's center of gravity and the others', showing divergence or convergence trends over the time range of the data set. Target is represented by the X-axis.
- Distance vs Trend: indicates the level of competitive caution of the target vs others. X and Y axes are the distance and trends between the target's center of gravity and the others' gravity. Positive and negative numbers indicate if their center of gravity is moving away from (+) or approaching (-) the target. Also see synergy classifications below.
- Distance vs Radius of Area: shows the overlap and distances between the target's patent distribution range and center of gravity compared to others in the data set. X-axis shows the area of distribution. Y-axis shows the distance between the target's gravity and the others. Plots to the left of the green line indicate an overlap with the target, while to the right of the line indicates none. Also see synergy classifications below.
- Number of Applicants in documents: indicates the number of documents an applicant has in the data set.
- Number of Inventors in documents: indicates the number of documents that an inventor is connected to in the data set.
- Number of documents in IPCs: indicates the number of documents containing a specific IPC.
- Number of documents in Applicant: indicates the number of documents an applicant has over the time-period of the data set.
- Number of documents in Inventor: indicates the number of documents that an inventor is connected to over the time-period of the data set.
- Number of documents in IPCs: indicates the number of documents containing a specific IPC over the time-period of the data set.
Contour Graph: shows the density of patents in areas of the radar.
- Levels: sets the number of contour line intervals. The higher the number the sharper the contour regions.
- Range: controls the displayed portion of Levels.
- Factor: increases or decreases contour line detail in low density areas.
- Reset: returns the settings to the default.
- Create Areas from Lowest Contour Lines: shows keywords or lowest contour levels, and activates Areas functions.
- Update the Contours Automatically: adjusts the contour regions automatically based on Highlight selections.
noteIf the Contour Region has a square outline, it can be removed by lowering the Factor Level, or increasing the Contour Range Threshold in Advanced Settings - accessible through the gear icon (⚙) in the bottom right corner of the radar. [Visualization → Advanced Settings]
Trend Graph: shows the target and its competitors’ centers of gravity movement over the time range of the data set. Click on a company to view its trend graph, and single click on the circle next to the company name to change the highlight color.
noteYou can add or remove companies on the list by clicking the gear icon (⚙) on the bottom right of the radar
Quick Highlights: highlight or isolate clusters based on keywords, years, assignees, inventors, or IPCs.
OR displays the clusters that contain the selected items.
AND displays the clusters that contain ALL of the items selected.
FILTER displays ONLY the selected items.
SCALE resizes the clusters based on volume relevancy of the selected highlights. Single click on the circle next to the selected items to change the highlight color.
Pins: contains a list of pins you have placed on the radar. Click the pin icon on the radar to add or edit.
Comments: contains a list of comments you have placed on the radar. Click the comment icon on the radar to add or edit.
Filters: create and manage customized filters.
- Name: enter filter name (required).
- Color: select color.
- Border Width: set 0 for a cluster fill. Set 1-10 to circle the cluster. The higher the number the thicker the circle.
- Shared Filter: set ON to make filter shareable across all analyses.
- To create a filter: select the filter target → select the target operator → enter the condition value.
- To apply multiple values: fill in fields of Condition 1 and Condition 2 (optional).
- Save: click save to create and apply the filter.
Condition Option - Notes:
- Data sets: enter the number located on the data set details pop-up on Data Sets page.
- F-Terms: applicable only with Japanese patents
- Cluster ID: enter a numerical value
- Live or Die: terms for Application status:
- Patent Office and Status values:
- JP Application → jp
- JP Granted → jppub
- US Application → usapp
- US Granted → usreg
- EPO Application → epa
- EPO Granted → epb
- WIPO Application → woa
To view, edit, and delete shared filters, use Manage Shared Filters.
Areas: select and analyze areas on the radar.
- Click Create Area icon to activate the function and then draw on the radar to enclose the area you want to analyze.
- Click Select Area icon, hover over a contour region and single click to select areas by contour line.
- Click Select Area icon, hover over a selected area and single click to view documents in area, edit area information, create a data set from area, create a sub-analysis from area, or delete the area. Or, view and manage areas on the Areas list.
Center of Gravity and Distribution Area: shows the overlap and distances between the target's patent distribution range and center of gravity compared to others in the data set. Single click on the circle next to the selected companies to change the highlight color.
Text Plots: places an X on the radar to indicate the center of an area containing specified text. Enter single or multiple words in the search query and click save to mark the spot.
Drawing on Radar: draw and label areas on the radar.
Radar Settings: adjust the size of the clusters.