5.1.1. Carrot² [CAR 17]

This is one of the most well-known specialist search engines; it uses the Web and Medline [PUB 17] as sources of information, but also uses various Wikis [WIK 18b] (free access to medical and paramedical publications and to explanations on various Wikipedia pages).

Figures 5.1, 5.2 and 5.3 show the different results obtained from a question (Moringa [WIK 18c])² and the sources of information above in the form of a bubble diagram. The results are shown as interactive bubble foams. The size of a bubble foam indicates the importance of the Websites involved. By selecting a bubble foam, we reach different sites which can then be consulted directly. We can also use other representations or even simply use the clustering used in the foam diagram.

Here, we have tackled a general topic, but it is possible to work in the same way using the names of authors, societies etc. The results shown here were obtained on February 8, 2017.

5.1.2. Wikimindmap [WIK 18d]

This application is provided here for information, to demonstrate the creation of heuristic maps. In fact, after ceasing to function for a time, this application reappeared on the Web, but no longer seems to be available. However, an API is available to integrate this into clustering applications [API 17].

Here, we use various Wikis as sources of information, but not always in general, they are chosen according to the language used by the Wiki. Notably, this makes it possible to see how a single subject is treated in different countries. It is the English Wiki that will lead to the best results, as it contains more subjects. We always use Moringa as the query.

We note here the difference between the two types of Wiki (Figures 5.4 and 5.5). The representation of the French Wiki is richer and this is certainly due to the fact that Moringa is widely used in French-speaking areas of Africa. Also, note that one can click on the plus (+) signs present after the data to extend the map. When two green arrows follow, this means that one can access a website by clicking on them.

5.1.3. Newsmap [NEW 17]

At the time of writing (February 2018), this site is not accessible and we do not know if it will return. This specialist search engine works mainly using print media. It is generally used for geopolitics, because one can choose a country of origin and a domain and visualize different articles on the subject. The larger the rectangle, the more data there is available. When it is used for free, one can only access one article per domain (per rectangle containing information). It will be noted that on the bottom left of Figure 5.6 there is the date: February 8, 2017 and a choice of analysis time: 10 minutes, more than 10 minutes, more than an hour. On the upper row: the choice of countries, on the lower row, the choice of themes: world, national, business, technology, sports, entertainment, health, etc. There also seem to be some difficulties with accessing this application.

On the theme “the electro-magnetic tether for disorbiting satellites is a failure”, there are 13 articles on the subject, but the free version only allows access to one. An extract of this is given in Box 5.1.

There are other specialist search engines, such as Tagcrowd [TAG 18] which selects the most significant words from a text and represents them in varying sizes depending on their frequency, while iBoogie [DIR 18] which is a specialist engine for searching in preselected commercial searches, but which is no longer accessible, etc. In this domain, applications emerge and then, just as quickly, disappear. Here, we have indicated the most “stable”, which have lasted over time.

5.2.1. Google Scholar

Generally, a search on Google Scholar leads to a succession of data presented according to the search characteristics of the Web. This data cannot be downloaded in “packets”. Its benefit lies in the fact that Google Scholar has one of the best coverages of the sciences, as very diverse sources of information (all of scientific nature, including American patents) are indexed. This makes it possible both to make a fairly exhaustive search and to access documents in different languages (depending on the language used for the enquiry), but also to access data formats for the data sought, which is particularly useful. Here is an example of information obtained from Google Scholar, which gives access to information not found in classic databases.

Clicking “Cite” (see below) provides access to different citation formats:

• – Joachim J., Kister J., Bertacchini Y., Dou H., Intelligence économique et Système d’information, 2006;
• – Joachim J., Kister J., Bertacchini Yann et al., “Intelligence économique & système d’information”, Revue ISDM, vol. 24, 2006;
• – Joachim J. et al., “Intelligence économique & système d’information”, Revue ISDM 24, 2006.

NOTE.– [PDF] means the whole text can normally be accessed for free.

J. Kister, Y. Bertacchini, H. Dou: the authors’ names are underlined, which means that their profile can be accessed in Google Scholar. An example of a profile obtained by clicking on an underlined name is shown in Figure 5.9. To create a profile, you need to register (for free) in Google.

5.2.2. Access to Google Scholar since PoP (Publish or Perish)

An original form of processing information obtained from Google Scholar has been developed by Anne Harzing [HAR 07]. She developed a software (PoP) that makes it possible both to extract data from Google Scholar (up to a thousand references), then to structure it and present it in table form. At the same time, a number of factors (h-index, impact, etc.) linked to the data’s statistical processing are also shown. It is possible to search by theme and by author name, which is a quick and general approach to covering a domain or author. The software used can be downloaded for free and of course, access to the source of information is also free. Figure 5.10 shows the result of a search on the topic: “competitive intelligence” and “strategic dependence” between 2000 and 2017.

Results can be saved in text, CSV or excel format. In this example, we will not expand on the meaning of the different indexes [HIR 17]; these relate to data on the number of works published by the authors, the number of citations, etc.

It is also possible to save all references in different formats. This is practical, as some of these saved files are formatted, which makes it possible to process them easily using bibliometric analysis software [MAT 17c]. Although direct access is not allowed by Google Scholar, using PoP as an intermediary it is thus possible to access data sets (no more than a thousand) taken from Google Scholar and formatted.

5.2.3. The Web of Science (WoS)

The Web of Science is a multidisciplinary database which is often used at a scientific level to evaluate researchers. However, it is critiqued by the fact that it has relatively specialist coverage (major scientific publishers, preponderance of the English language and selected journals). The three databases most often used – Web of Science, Google Scholar and Pubmed – show significant differences that have been analyzed by the Michigan State Library [MIC 17] in the United States:

“Both Pubmed and Web of Science are human-curated databases. Google Scholar is not. This is the key to most of the differences you will find in your search results […]. Google Scholar is not a human-curated database but a search engine of the whole internet […]. Both the Web of Science and Google Scholar take care of citations, Pubmed does not […]. Google Scholar searches full text of articles but PubMed and Web of Science search only the citation, abstract, and tagging information”.

Enquiries in the WoS are made in the classic way and results are downloaded in the form of a text file, which is naturally formatted. For analysis, we then use (from tens to several thousands of bibliographic references if necessary) the Matheo Analyzer [MAT 17c] software, which is one of the best compromises between performance, cost, speed and ergonomics. In the example shown below (Table 5.1), a search has been made on WoS with the term Moringa, between 2010 and 2014³, which gave:

Which shows major interest in products coming from this plant from 2011. The enquiry returns:

We can then recombine information according to needs. For example, creating a network of main actors, with the following parameters: form (author) frequencies 5–17, Pair frequencies 1–17, Connectivity 0–66.

Beside each author the number of publications is indicated and on each link, its frequency is indicated. Where authors’ names aren’t linked, this simply means that the author has published alone, since the frequency threshold chosen for the links between authors is 1.

Figure 5.12 shows the emergence of concepts over years of publication for the concept “coagulant”. It can be shown that this concept only gained momentum from 2012.

One can also compare different concepts with sources of publication, laboratory addresses, countries of origin, etc. This makes it possible to have an overall view of the “subject” Moringa at the level of areas of work, but also the countries and laboratories involved.

5.2.4. Pubmed

This database is free and any enquiry result can be downloaded in structured text format. Importation into analysis software is immediate, as the importation table is native. The search using the term Moringa was carried out between 2013 and 2015. We obtain:

We can then combine the different extractions to obtain meta-data providing general information on the subject. In the following figures we find some examples.

Figure 5.15 represents inter-university collaborations. The matrix shown (partial view only) is a squared matrix made with different publication affiliations. The intersection of the columns and lines gives the number of publications made through collaboration with two universities. The names of the universities are abridged, but by moving the mouse over the abridged name you can see the whole name.

The data extracted can also be used as a powerful search tool, by comparing, for example, the titles (it is possible to select a word in the title, here “water”) and obtaining sources of information afferent to the titles selected. In Figure 5.16 we show the comparison between the “concept title: water” and the authors. The concept title is created automatically by analysis software.

In this presentation of an analysis of the information obtained from scientific publications, we have not drawn any conclusions on the types of work and results. We have simply shown the extent of the correlations that can be made. The choice of correlations and their interpretation lies to the user and their needs.

We note the multiplicity of sources of information in particular, since in the selection made in Pubmed, we have publication by source, which leads to a very substantial dissemination of the results.

5.3.1. General remarks on patents

A patent only protects an invention in a country where the patent has been filed.

A patent can be extended after it is first deposited (priority date) in other countries. There are diverse procedures for doing this: using national patent offices, in this case, there is a wait of a year after it is first filed.

One can also use PCT procedures (global patent WOxxxxxx), or European procedures (European patents or EPxxxxxxx). If, after examination, the patent is granted, it is then sent to national procedures to determine which States it will be protected in.

So, a single invention can be covered by patents with different numbers that will then form a family [EPO 17]. When only one patent is not extended, the family only contains one member.

A patent’s lifespan in 20 years. A five-year extension is possible for some patents in the United States [WES 17], and for pharmaceutical products in Europe, it is possible to prolong the patent’s lifespan [MAR 97]: the Supplementary Protection Certificate (CCP – Certificat complémentaire de protection) is a title extending the legal duration of a pharmaceutical patient by the period necessary for obtaining market authorization, a lengthy and complex administrative procedure aimed at protecting human health.

Currently, China deposits the most patents in the world. Chinese patents can be accessed in English via the CNIPA site [CNIPA 17] (National Intellectual Property Administration), but the worldwide database [WOR 17] (Espacenet Patent Search) also makes it possible to access Chinese patents easily. Moreover, it has summaries of utility models which is not the case with the CNIPA database. Around half the patents deposited are utility models [BEC 13, STA 10]. The utility models are “petty patents” which are easily granted (there is simply an examination form and they cost much less than a patent), but which can still be opposed by third parties (it is at this moment that a substantive examination is carried out). The lifespan of a utility model in China is 10 years but it may be less for other countries. Although in the west, we consider utility models as relatively unimportant, it is vital for businesses working in, exporting to, or filing patents in China to be examined. The case of Schneider Electric is a good example, as failure to take account of a utility model forced this firm to pay a fine of 23 million US dollars to a Chinese business:

“The effectiveness of utility models was demonstrated in recent litigation in China between Chint, a manufacturer of low voltage devices in China, and Schneider electric. In that litigation, Chint succeeded in asserting its utility model against Schneider and obtained a verdict of $49 million, a productive result for a utility model application with just a $70 filing fee. Chint’s utility model survived Schneider’s subsequent invalidity challenge, with Schneider settling the lawsuit for $23 million. While the dollar amounts might seem small in comparison to patent litigation in the U.S., the Chint case has been described as one of the largest patent damages verdicts in China to date. When to use a Utility Model there are a number of situations when a company should consider filing for one or more utility models” [JEW 10].

Some national patent offices, for lack of means to examine patents, consequently deliver patents on formalities and not after substantive examination. In this case, extending these patents at international level (WO, EP, US, etc.) where substantive examination is carried out carefully is not often possible. It is therefore necessary, if we wish to use one of these patents, to verify its “solidity” at the level of scientific and technical content. If they are under examination (US WO EP procedure), the Patentscope database [PAT 17a] provides information on the state of examination and indicates the patents judged to be enforceable by the examiner. If patents are classed as X or Y [BRE 17], this indicates that opposition is particularly strong and that the patent examined will not be granted, or various claims will not be upheld.

In addition to fees for deposits and specialist consultants, protection is kept up to date by the payment of annuities, which will increase as the patent ages. This is why it is generally said that patents more than 10 years old are often abandoned. We can verify this using specialist databases. However, there are no systematic means of knowing if a patent is in use or not.

Of all the patents deposited, it is estimated that only 20% are actually used. This figure varies depending on the country, for example in Algeria only 0.1% of patents are used [ALG 17]. Some authors even indicate that 97% of patents deposited do not bring in any money [KEY 17]. We can also consider the cost of depositing patents whose content has little chance of being accepted. Often, this situation results from very poor documentary research at the outset. This can be carried out internally by the depositing party themselves, or we can then have recourse to what is now called “the Virtual Patent Office” or VIPO, which is in fact a specialist research bureau that – for a considerable cost – will carry out research for you. Thomson Reuters announced the following figures for depositing US and EP (European) patents:

“Hundreds of thousands of patent applications are filed every year in the US and Europe (brevets européens). Over 500,000 patent applications are filed per year in the US, but only around 300,000 granted patents are issued. Similarly, about 150,000 patent applications are filed every year at the EPO, but only around 60,000 granted patents are issued. Globally, there are over one million patent applications filed annually but only 50–70% of them result in a granted patent” [RES 17].

This therefore results in:

“A very conservative estimate for the cost of drafting and filing either a US or EP (European patent) application is around $ 10 000. Based on that estimate, the 200,000 applications that are filed in the US per year that will not result in a grant adds up to 2$ billion. For the EPO, it comes to another $ 1 billion spent annually on drafting and filing applications that will not yield a grant. For the US and EP alone, this is a combined gap of $ 3 billion per year.”

You should therefore bear these statistics in mind if you wish to deposit a patent, first be sure that you or someone else will be able to use the invention and above all, be sure that a study of the previous state of the art shows that there is a good chance the patent will be granted. One should also bear in mind, that although we are straying into a legal framework here, that litigation over intellectual property is usually inordinately expensive.

If a patent is not used, there is legislation in France on lack of use [INP 17]:

“The legislator can impose upon the owner of the patent the obligation to exploit it. A patent is considered unused if three years from delivery of the patent, or four years counting the date the request is made, the owner of patent:

• – has not begun to use the invention that is the subject of the patent, nor made effective and serious preparation to do so;
• – or if they have not sold the product that is the subject of the patent in sufficient quantity to meet the needs of the French market;
• – or if they have abandoned the use or marketing of the patent for three years.

Where the patent is not used after this time span, the sanction is that a license must be granted to any other individual who requests it.

To make a request for a license under these circumstances, the individual making the request must show that:

• – they are able to use the invention seriously and effectively;
• – and that the owner of the patent has not granted them license to use it.

The High Court tribunal, which the delivers the obligatory license, attaches conditions to it: duration, field of application, amount of royalties”.

5.3.2. Analyzing patent information

The examples we are about to show are made from two specialist software (Patent Pulse and Matheo Patent) that allow:

• – remote work. Direct processing from your computer, considered as a terminal through an interface accessible via the Internet [PAT 17]⁴, the processing, analysis, data extractions and queries exported to a distant server. The following databases are accessible: Worldwide (EPO), USPTO Published full text (United States), USPTO granted full text (United States), EP full text, PCT full text (global patents), FR full text (France);
• – local work. This means downloading data onto your computer followed by local processing on a resident specialist software [MAT 17b]. The following databases are accessible: Worldwide, USPTO Published, USPTO granted.

Here, it is not a question of tackling the analysis of patents (APA Automatic Patent Analysis) in detail, but of alerting the reader to their importance and showing the powerful processing treatments making it possible to explore subjects in an unconventional way.

5.3.2.1. Remote processing

The characteristics of the software used have been described in the literature [DOU 16d]. This working platform (Patent Pulse) makes it possible to carry out searches in different databases, to save requests (remotely) or references downloaded (each time a request is made, the search is updated, while if references are stored, it is for the user, if they wish, to update their data). When the data are downloaded, automatic access to different histograms is suggested. The following documentary fields are involved: country (i.e. national) codes, depositors (applicants), the inventers, the international cooperative classification, the international classification, the US classification, the date of publication, the application date, the priority date.

The user can then make histograms, matrixes and networks, using the following fields: depositors, normalized depositors, country of application, inventers, inventor’s country, CPC [WIK 18f], CPC4, CPC7, IPC [INT 17], IPC4, IPC7, US class, US main Class [USP 17], JPFI, JPF Term (these last two fields involve a classification for Japanese patents [JAP 17]). Thus, to quickly compare the inventors’ levels of expertise, an inventor matrix, IPC or IPC4 for example, will provide a comparative “bench mark”. The depositors can also be treated in the same way, etc. A filter also makes it possible to carry out searches in the downloaded base.

Finally, data analysis should be completed via interaction between experts, to finalize more complete and exhaustive analyses and vary cognitive and economic approaches to the subject. Ronald N. Kostoff [KOS 01], in various documents, describes how using literature analysis through bibliometry can hasten discovery and radical innovation in science and technology. In a more specific study, he shows that to stimulate participation in workshops, the use of email beforehand was crucial [KOS 02]:

“The e-mail component of the workshop is crucial. The gestation period between inputting promising ideas and real discussion during the workshop makes it possible to consider many different approaches and syntheses. It also makes possible to save a lot of time in the workshop by clarifying confusing questions at the outset”.

Other studies also show that interaction between experts is a key stage for stimulating ideas and makes it easier to come up with new ones. The concept of “ba” [NON 00], Japanese for “space”, where interactions can emerge, underlines the importance of a physical or virtual space where experts can develop “alerts” (or in the best case “key knowledge topics” KIT). Nonaka, one of the authors, explained:

“What differentiates ba from ordinary human interaction is the concept of creating knowledge. It is from such a platform that the transcendental perspective integrates all information needs”.

The concept of platforms for developing different types of knowledge, and then the competitive advantage, has been extended to many domains such as regional development [ASH 11]. One of the main points, underlined by Purvis [PUR 01], is assimilating knowledge platforms into organizations:

“Nevertheless, to reach the highest potential of organizational knowledge, we must not just be content with adopting and using “IT-enabled” knowledge platforms. These platforms should be assimilated into the current working process within organizations”.

It is within this framework and by taking account of previous considerations that a system for analyzing patent information was set up, allowing users not only to select different sets of patents and carry out different analyses, but to create with different users, KITs (Key Intelligence Topics) depending on the importance of the subject or its possible evolutions. The system used here, accompanied by notes, their transfer to other users, the sharing of indexes containing the notices downloaded, etc. were designed to breakdown the user’s potential isolation. The granting of multiple licenses enables intra-company work but it is also possible to work with other users outside the business, by generating all the exchange links created confidentially.

We will therefore present an example of processing that will complete the classic approach to making queries in a patent database. Generally, the questions asked may combine the previous fields (cited above in the example tackling the creation of histograms or matrices) plus the terms used either in title or in titles and summaries, or in the whole text when it is accessible. This “classic” approach makes it possible to select a set of patents whose content will then be analyzed. In our example, a search with the term “snowboard” creates to a database, part of which is shown on Figure 5.17.

The patent shaded in Figure 5.17 is of particular benefit for the user. Figure 5.18 shows its bibliographical notice.

We can see that the patent number is followed by a letter and a digit. This letter and digit indicate how far advanced the patent examination is [SIG 17], that is its status. For example:

For American patents, consult the meaning of the kind codes (name given to the letters and digits above) [KIN 17]. In the present case: A1 = Patent Application Publication.

This patent is part of a family of patents covering the same invention, this is the first point. It is also present in the downloaded database, issuing from the request “snowboard”. But, are these two sets enough to be sure that the majority of patents afferent to the theme of research have really been identified? This is not necessarily true. To partially answer this question, we will, starting with this patent, generate a group of similar patents by creating meta-data from its citations. In fact, patents, especially WO, have the peculiarity of citing patents and being cited by patents. There is no question here of addressing all possible analyses of citations [LIX 07, MIC 01], but of giving an example to give the economic intelligence practitioner a broader overview of the use of citations.

We therefore click on the citation window (upper band), and we obtain the result shown in Figure 5.20.

We will now focus on the patents cited and especially on patent US66739615B1. To do this, we “open” the citations for each of the patents cited below (select the patent with the mouse, then click on the icon on the top left, which opens the citation space). This is done for each of the patents shown on Figure 5.20, until we stop at the American patent, which is different.

This patent is key for us, as it is cited by many other patents and especially by the patent that is the focus of the user’s interest (WO201607744A1). We can thus conclude that the set of patents that cite this one (a very large set) are, because of this citation, directly linked to patent WO201607744A1. Because of this, all these patents are potential candidates as subjects of interest. We then automatically recover the numbers of these patents, to form a database of them automatically, which is then downloaded. This set of patents, some of which may be present in the initial database on “snowboard”, broaden the initial research, which is not based on a classical combination of Boolean operators. This base is then analyzed in detail and the results incorporated into the user’s concerns. This type of broadening of research is a holistic example of how to use meta-information.

5.3.2.2. Processing carried out on a local database with an appropriate analysis software

Here, we will address a simple subject, with local processing of the analyses. A firm is interested in an analysis of welding equipment. We carry out research using the global patent database, with the terms “welding apparatus” present and patent summaries and titles, for the period 2009–2012. The period does not much matter for examples of processing. We thus obtain 3,211 patents distributed across 964 families. From this research, we will introduce as an example the search for “strategic dependency” [DOU 09, DOU 12] for some countries for the domain considered. In the same way as for previous analyses, the set of references is broken down according to the needs of the user.

Number of patents	3,211 distributed over 964 families
Concepts: types of welding that interest the user	14 (the data are indicated by the number of families)
Friction	91
Ultrasonic	43
Pulse	18
University as depositor	13
PR=US Number of US priority patents	122
PR=CN	86
PR=KR	273
PR=JP	397
PR=DE	25
PR=EP	32
PN=US Total number of US patents including extensions	257
PN=CN	265
PN=KR	337
PN=JP	429
PN=DE	75

The concept of strategic dependence is introduced in the following way: a patent is first deposited in a given country, generally the home country of the depositor or one of the depositors. Then it (patent N) can be extended following either national procedures, or by it becoming a European or global patent with a choice of states where the invention should be protected (states O, P, etc.). Thus, it is possible to determine the contribution of a patent N to the strategic dependency of this patent in states O, P, etc.

Patent WO2012041375A1 (“A welding apparatus and a method for welding”) is a European patent of Swedish origin, which was extended to the United States following a PCT procedure. Thus, the United States display a strategic dependency on Sweden for the topic the patent addresses, which is a type of machine for welding.

This means that, if the patent has been accepted, the following states will have dependency where this technology is concerned: China, Russia, Japan, Australia, Canada, South Korea and the United States. In fact, they will not be able to use the invention described freely and they will either have to create a partnership with the Swedish firm, pay them royalties or reach a licensing agreement. This is the concept of a country’s strategic dependency where a particular technology is involved. In the present case, for all the patent families that fall under the topic “welding apparatus”, we can calculate the strategic dependency of various countries on other countries (or businesses). To do this, we will create the following matrix.

The matrix can be read line by line:

• – those that have been extended to Germany (DE): 18 Japanese patents, 29 American patents, one Chinese patent and two South Korean patents;
• – those that have been extended to the United States (US): 94 Japanese patents (including six with double Japanese and American priority), two Chinese patents, four South Korean patents and 16 German patents;
• – those that have been extended to China (CN): 106 Japanese patents, 48 American patents, seven South Korean patents, nine German patents + four (EP), 10 EP patents (including one French, four German, three Swedish and one where the depositor is not given;
• – those that have been extended to Japan (JP): 33 American patents, two Chinese patents, two South Korean patents and seven German patents;
• – those that have been extended to South Korea (KR): 41 Japanese patents, 22 American patents, two Chinese patents and two German patents.

Some results show a very low variation compared to manual counting. This is due mainly to patents that show a double priority and to some European patents that simply have priority in Europe, that is the European patent. In this case, we must return to the depositor’s nationality to determine their country of priority.

This method of working, which calls for the use of meta-information, is particularly powerful. It could be used when deciding on a region or country’s strategic technologies or to decide a policy of extending patents or licensing in start-ups or universities. Knowledge of a country’s technological dependency in certain domains should be measured to guide investments in domains where blocking due to foreign patents is not evident.

We will now briefly address the concept of new entrants. The five Porter forces show a diagonal formed both by the appearance of new technologies and by the arrival of new entrants in the domain considered, these entrants are still unknown and may disrupt the market. Analyzing the information contained in the patents is one instrument of choice for finding new entrants or detecting the appearance of new technologies.

This last point is fairly easy to address, as starting with the general subject “welding apparatus”, automatic access to words contained in the titles and summaries as well as the creation of “concept titles” makes it possible to detect new technologies or directions simply from reading. It thus makes it possible to find the patents that describe them and regroup the information into a general concept that will then be analyzed. On the other hand, detecting new entrants will require the creation of a general matrix of priority dates versus applicants. The priority data is always used as the date, because it is concrete and not subject to variation [POR 07]. Applications are described in Figures 5.24 and 5.25.

In this partial view of the matrix, “priority years versus applicants”, the potential new entrants are businesses that did not have patents in previous years and which only appear in recent years: Amaperex Technology Ltd. only appears in 2011, the same goes for Ningde Amperex Technology Ltd or Seidensha Electronics, etc. On the contrary, some applicants no longer appear in recent years and have thus disappeared from the field of activity (at least as far as intellectual property is concerned). The same process can be carried out for each of the themes selected. One can also, if necessary, evaluate the novelty of these topics by comparing the latter with the priority.

At a topic level, we see that friction welding is a fairly constant activity, ultrasonic welding is in relative decline and the domain of magnetic pulse welding, although it appeared quite recently, also seems to be disappearing rapidly.

For all patents – including priorities and extensions – we see that the main activity in this domain was focused in the years 2006 to 2011 and that afterwards, there was a noticeable drop in interest.

We note that the method’s reliability is linked to the quality of the query. If through lack of vocabulary we “lack” some significant patents, it is clear that the final results are wrong.

This way of working, which gives a global view of a domain’s evolution, enables global reflection and positions the paths to be taken in a broader and more international context. In this sense, it is very useful for making dashboards.