También disponible en Español

Inf@Vis!

The digital magazine of InfoVis.net

Patent Analysis
by Juan C. Dürsteler [message nº 167]

Patents are extremely important when configuring the business strategy of technological companies. There are more and more tools that allow you to download and analyse patents in one way or another. Visualisation is a key tool for the analysis and detection of opportunities that is still used shyly.

One of the less usual applications of information visualisation is patent visualisation. In fact visualising  patent information is just a part of text or document visualisation. In the end, a patent is just text with a certain structure and with a specific objective. 

Patents encompass several aspects that have, nonetheless, great importance.:

  • they are a way to protect intellectual property rights

    • our own rights, allowing us to produce without being copied with inpunity

    • the other's rights, preventing us of using alien developments. 

For these reasons patents constitute a very important source of information about the competitors. This makes them the main tool for technological and business intelligence. On the other hand, knowing which patents are enforced in the market is vital before deciding to make a new development or product, with the important investment it means, that could be finally blocked by an existing patent.

Many patent offices already allow to freely download abstracts and complete text of their patents. Among them the USPTO from United States or the EPO from Europe, and many others you can find, for example, in the PatentLawLinks.com list. In particular Esp@cenet, the download service of the European Patent Office has become a very popular source of information in this field.  

Based on the many possibilities of search that all this offers, there have been appearing in the market different systems allowing you to search, download and analyse patents automatically. Many of them are present in the Patent Information Users Group (PIUG) vendors list. Although downloading, classificating, clustering and finding the relationship between patents of similar contents has become quite widespread, even making use of text mining tools, the visualisation of those resuts is not so usual.

Anacubis, patent division of the i2 inc group, has some visualisation demos that use  i2 inc Analyst's Notebook visualisation tool. Unfortunately some problem with the demo has prevented me from executing it properly, so I can't report on it.

Spore.inc provides two main types of visualisations: 

Patent matrix diagram. Diagrams that represent the substance and hierachical relationships of the claims of a patent. This way, reading and understanding of a patent's claims is much  easier.
Source: Diagram as can be seen at Spore Inc website.
Spore diagram. It produces graphs of patent groups that allow you to see how are they related, identifying trends and detecting development opportunities related to "gaps" in some areas of the patent scope of a product portfolio.
Source: Diagram as can be seen at Spore Inc website.

But maybe Mathéo Software represents one of the easiest to use systems (you can download a free demo) that incorporates four main types of visualisation that can be produced with many combinations of the different variables that identify a patent. We'll focus on MatheoPatent 6.1. This program allows you to launch a search on different sources according to keywords, inventors, etc. With the downloaded results you can get the following types of visualisation.

MatheoPatent 6.1

MatheoPatLista2.gif (182879 bytes) MatheoPatPDate.gif (68024 bytes)
Table: Presentation in form of a table where you can see the results of the search. To the left, a list of inventors with their nationalities and number of patents granted. To the right, the list of patents. Below lies the summary of particular patent of the list.
Source: Screenshot by the author of the program in execution.
Click on the image to enlarge it.
Bar chart: Here you can see a bar chart of the number of patents present per year. MatheoPatent allows you to select different variables (inventor, family, date,  etc) building bar charts accordingly. 
Source: Screenshot by the author of the program in execution.
Click on the image to enlarge it.
MatheoMatrix1.gif (146436 bytes) MatheoGrafo1.gif (82098 bytes)
Matrix: By crossing the different variables MatheoPatent allows you to select (inventor, family, date,  etc), you can obtain this matrix chart. In this particular case we cross inventors against inventors. The coloured squares in the diagonal indicate the patents of a particular inventor. Big squares indicate groups of inventors acting together. If they are out of the diagonal it means that some inventors or groups participate in patents with other groups. It's possible to cross other variables to detect different patterns
Source: Screenshot by the author of the program in execution.
Click on the image to enlarge it.
Networks: In this case we represent the relationships between inventors and companies by means of a force directed graph where dragging a company you can drag also the inventors related to it as if they were linked by rubber bands. Again we can select different variables to see the networking between them.
Source: Screenshot by the author of the program in execution.
Click on the image to enlarge it.

¿Do you remember ThemeScape and newsmaps? (see number 93). Finally this technology, renamed as Aureka!, has become the patent visualisation of MicroPatent. This way, this extraordinary form of visualisation has seen its incorporation to the world of patents. Let's recall that ThemeScape allows you to represent a document corpus as a topographic map where the "mountains" are associated with frequent terms (predominant "themes") that are as close as their concepts are similar.

Aureka! The patent version of ThemeScape represents as a topographic map the explored document space. Dots depict particular documents (patents) . The closer they are, the most related the patent topics. You can access the documents by clicking on them.
Fuente: Image of Aureka! as can be seen at MicroPatent's website in Internet. 

Patent visualisation is just one more within the possibilities of text mining. Nevertheless, as anyone working in R+D knows, detecting the patents that can prevent us to follow a research line or finding a "hole" where there's nothing patented can be fundamental for the business strategy of a company. Visualisation is vital to open a way through the web of legal text, [sometimes obscure] claims, and the large amount of data that represent the world of patents. We are still an a very preliminar stage in this field.

Links of this issue:

http://www.uspto.gov/   United States Patent Office (USPTO)
http://www.european-patent-office.org/   European Patent Office (EPO)
http://www.patentlawlinks.com/patoff.htm   PatentLawLinks website
http://ep.espacenet.com/   Espacenet EPO download center
http://www.piug.org/   Patent Information Users Group web site.
http://www.piug.org/vendor.html   Patent Information Users Group vendors list.
http://www.anacubis.com/   Anacubis website
http://www.i2inc.com/Products/Analysts_Notebook/default.asp   i2 inc. Analyst's Notebook
http://www.sporeusa.com/home/   Spore website
http://www.matheo-software.com/   Mathéo Software website
http://www.infovis.net/printMag.php?num=93&lang=2   Num 93 Two years later
http://www.micropatent.com/   Micro Patent website
© Copyright InfoVis.net 2000-2014