New Zealand’s largest inventor network: A glimpse of our innovation ecosystem

By Shaun Hendy 04/04/2011 15


We have been experimenting with ways to represent the inventor networks that we can extract from patent databases.  In this post, I will focus on New Zealand’s largest inventor network, as extracted from 30 years of European Patent Office (EPO) data.  The network gives us a glimpse of New Zealand’s innovation ecosystem.

NZ inv largest v3At the left is a network we have constructed showing the largest group of connected inventors in New Zealand.  Each red dot represents an inventor, and the size of the dot represents the number of patents on which that person is named. Inventors are connected by a blue line where they have shared a patent.

There are four hundred and fifty inventors in the network, and it links fourteen New Zealand companies:  Fonterra, A2 Corporation, Fisher and Paykel Healthcare, Genesis Research and Development, Wrightson Seeds, Biojoule, Sensortec, Arborgen, Protemix, Neuren Pharmaceuticals, Brainz Instruments Ltd, Dashfoot Ltd, Vital Food Processors Limited and Envirofocus Ltd.  A number of international pharmaceutical companies are linked together through this network, including Chiron Corporation, Cancer Research Technology, Xenova Limited, Proteotech, Pharmacia & Upjohn (now part of Pfizer) and Warner-Lambert (also now part of Pfizer).

Four of the Crown Research Institutes are in this network (Crop and Food Research, NIWA, Industrial Research Ltd and AgResearch) as well as three universities (Massey University, the University of Otago, the University of Auckland) and the Malaghan Institute.  There are also scientists from several of the Centres of Research Excellence:  the Maurice Wilkins Centre, the Riddet Institute and the MacDiarmid Institute.

This is a very interesting snapshot of New Zealand’s innovation sector. It links several of New Zealand’s largest companies to much smaller startups.  It links a company that manufactures advanced respirators to a company that sells seeds.  It links researchers from several of our major public sector research organisations to those in several of our most R&D intensive companies. It really does suggest there is an innovation ecosystem out there!

Largest Inventor NetworkIf we geolocate the inventors by the addresses listed on the patents, we can get an idea of the geographical spread of the network.  The image on the right (generated using Google Earth) shows that the network stretches from Northland to Dunedin.  It is truly a national network, although its heart is in Auckland.  Try it out yourself – the kmz data file is available here (and you can get Google Earth here).  Viewing the data file in Google Earth will enable you to zoom in on particular regions in detail.  You will be able to see the connections between individual researchers.

Here is a close up of Auckland and its inventors in the network:

Auckland Largest Component

So go take a look.  In the next few weeks, we will be releasing the full EPO patent map of New Zealand in Google Earth and also a map that traces our international linkages.

We are also interested in improving the usefulness of these maps — at the moment you will find both the inventors’ and the applicants’ names on the map, but we would like to add subject areas and other information.  Let us know if you have any ideas!

P.s.  I would like to thank Catriona Sissons for her hard work in putting these maps together.  She has worked with me since 2008 on patent analysis, but recently moved to Melbourne to try life on the other side of the Tasman.


15 Responses to “New Zealand’s largest inventor network: A glimpse of our innovation ecosystem”

  • Shaun, have you detected power-laws functional relations in the analysis of your inventor networks yet? A tell-tale signature of SOC (self-organized-critically) phenomena?

  • I guess that one question that has to be asked about preferential attachment if it allows class-jumping in general (ie, the class mobility at one time-instant changes as one jumps from one class to another, such as from lower to higher or vice versa). The way I interpret preferential attachment is that no class jumping is allowed, which is unrealistic, since we see businessmen going bankrupt all the time eg, the IRD is pursuing vigorously property developers like Auckland Dave Henderson , Christchurch Dave Henderson, Terry Serepisos and others to bankrupt them for tax disputes. If the IRD is successful we see these wealthy individuals jump from higher ranking class to lower ranking ones. As far as I have read about those businessmen, they started like everyone else, ie, very little but worked their way up. So, if the IRD is successful, then those businessmen will jump back to where they started. They many jump back up again at some future times.

    The point I’m raising, is that preferential attachment seems to treat the quantity of interest (in this case the quantity is wealth) as something that eliminate/shield losses from happening at all, which is a bit unrealistic.

    Also we see individuals started out with little but ended up being wealthy after a certain period of time, eg: Sam Morgan, founders of Google, etc,… These individuals have jumped from lower to higher.

    Here is an interesting article on wealth distribution which allows class-jumping and also the power-laws is being observed.

    “Why it is hard to share the wealth”
    http://www.newscientist.com/article/dn7107–why-it-is-hard-to-share-the-wealth.html

    Papers from the conference proceedings at Saha Institute, India , mentioned in the article above, has been published by Springer:

    “Econophysics of Wealth Distributions”

    If you check out the table of contents from the title above, you can Google them in which you can find those pre-prints available freely (mostly) from various authors who contributed.

  • Yeah, I agree – inventors can cease to be inventive for all sorts of reasons and indeed this is not accounted for explicitly in preferential attachment. Nonetheless, the fact that a simple model like preferential attachment can generate power laws means you have to be very cautious before concluding self-organised criticality is at work.

  • Hi Shaun, this is fascinating!

    Can’t wait to see your final patent map. Lots of people in the Ministry of Science & Innovation will be interested in this so I’ll send links around,

    Stephanie

  • Hi Stephanie – great! I had hoped there would be some interest from MSI. I understand there are some problems getting google earth installed though?

  • Shaun, reading your other article on zipf law, it looks like that you have found a universality there. That’s a very interesting finding.

    I think that your finding is similar to the findings (from econophysics literature) of universality of market dynamics. Of course power-laws are being observed.

    #1) “Stock Return Distributions: Tests of Scaling and Universality from Three Distinct Stock Markets”

    #2) “Comment on `Tests of Scaling and Universality of the Distributions of Trade Size and Share Volume: Evidence from Three Distinct Markets”

    There is also a paper of preferential attachment on the same site as well.

    #3) “A Generalized Preferential Attachment Model for Business Firm Growth Rates: II. Mathematical Treatment”

    There may be also other papers from Prof. Gene Stanley’s website that may be of interest to you, I reckon.

    “Econophysics Research: Publications”
    http://polymer.bu.edu/~hes/econophysics/

    Are you submitting your work into a journal for publication? If you are, which one? You don’t have to state it here if it is something that you don’t want to share.

    Anyway Prof. Gene Stanley from Boston University Physics Dept (that link shown above) is the Co-Editor of joutnal, Physica A: Statistical Mechanics and its Applications published by Elsevier and he would be definitely interested in your work. Physica A is where most of the econophysics publications are being submitted to for publications, although some of them appear in Nature or Physical Review Letters and others. Just a thought.

  • Also the paper, “Zipf plots and the size distribution of Firms” from Gene Stanley’s website may be of interest too.

  • Thanks for the links Falafulu. Yes, we have been working on and off towards a paper for a while – we were aiming at Phys Rev Lett/E – but as it is not our day job it is taking longer than usual …

  • Shaun, may be you want to explore the following if it is applicable to your inventor network. I think you already work with (or know about) “Ising Model” in statistical-mechanics for ferromagnetic spin-flip, which is what the paper is based on.

    #1) “Random field Ising model and community structure in complex networks” appeared in The European Physical Journal B – Condensed Matter and Complex Systems
    Volume 50, Number 3

    Also explore the following on “Minority Games” (MG) if it also applicable to your inventor network. The derivation of MG is based on “Icing Model”.

    #2) “Theory of networked minority games (MG) based on strategy pattern dynamics” appeared in Phys. Rev. E 70, 2004.

    Perhaps MG can be applied to find out if there is emergent behavior in the network (small or big), such as cooperation or competition amongst the agents/nodes or in your network, the inventors amongst themselves.

    MG as it was first proposed in the late 1990s is quite popular in econo-physics as a model to based (further derived) one’s trading strategy on. I’ve read on the net that some financial institutions have linked up with researchers who specialize in MG. There is no surprise there as everyone is looking for some winning strategies.

    Interestingly, I came across the following paper on the application of “Ising Model” in functional brain networks.

    #3) “Ising-like dynamics in large-scale functional brain networks” appeared in Phys. Rev. E 79, (2009)

    Freely available pre-prints of the above publications can be found in the link below.

    http://arxiv.org

  • Here is a nice article on minority games (MG), targeting general readers.

    “Minority Games”.
    http://news.softpedia.com/news/Minority-Games-38625.shtml

    The great surprise about the application of MG to economics is that it has reproduced all the statistical observation facts in economics/finance called stylized facts. The surprise is economists have always described stylized facts as statistical observation but puzzle why? MG puts stylized facts right on the doorstep of complex system in what a true economic system should behave. An economic system must be treated as a complex system (holistic) rather than from Micro/Macro-economics which is based on the single rational agent theory (single component) applied to the system (ie, network of connected agents – traders, firms, financial institutions, bankers, etc..).

    Economic representative theory cannot explain market crash or bubble although we see them all the time in the real world, but MG model and similar ones from statistical physics have managed to explain their existence, ie, running (simulation of) these models showed why market crash, bubbles. They also revealed emergent behaviors in the market place (true complex system dynamics), ie, cooperation & competition in free-markets, etc,…

    Its amazing how a concept about “ferromagnetic spin-slip” can be found (or derived other models from) to describe complex system.