Imagico.de

blog

Science and OpenStreetMap and why not call it ‘VGI’

| 11 Comments

These days OpenStreetMap is getting quite popular as a subject of study in social and geoinformation sciences. A lot of research papers are published on this matter and the term ‘VGI’ (volunteered geographic information) is a popular buzzword in this context, it seems to become almost a fixed requirement to use it in any scientific text dealing with OpenStreetMap in some form. Nearly always it is essentially used as a chiffre for OpenStreetMap – the idea being that by genericizing you make your results more universal. This almost never works since either you are only dealing with OpenStreetMap and not with any other alleged ‘VGI’ projects or you are putting OpenStreetMap in a box with other very different projects of crowd sourced data and are probably neglecting to properly analyze the differences.

On its own the term ‘VGI’ is a really bad choice due to several problems:

  • The key term is ‘volunteered’ which bears the question what it is meant to contrast with, i.e. what ‘non-volunteered geographic information’ is supposed to be. Also the question arises if and to what extent the data in OpenStreetMap can actually be considered volunteered more than data in other geographic databases. It if fairly unclear who volunteers what in case of OpenStreetMap – is data mapped by a paid mapper based on satellite imagery ‘volunteered information’? Is it volunteered by the mapper? By the satellite image provider? By the owner of the satellite?
  • The use of the term ‘information’ emphasizes these problems even further since it specifically refers to the underlying semantics of geographic data and not the concrete data representation.
  • For someone to actually substantially ‘volunteer information’ this would in fact have to be private, privileged – like when you map details of your backyard or inside your house only accessible to you. One of the core principles of OpenStreetMap is however verifiability – so information is either not really suited for OSM or it is not really volunteered since it is openly accessible to others as well so it is at most voluntarily entered by the mapper into the OSM database (although again the question would be how this could be non-voluntarily). So in a way this term is quite demeaning for OpenStreetMap since it does not acknowledge the principle of verifiability.

Although i am not sure if this played a role in coining this term ‘VGI’ underlines the idea that information is property and can and is usually owned by someone and that such information can only become free information by the owner volunteering it. This is actually diametrical to the core idea of the OpenStreetMap project to collect information in an open database that is inherently free because it is independently verifiable by anyone.

So if you want to use a catchy term something like ‘crowd sourced geodata’ might be more fitting but you’d better think well if it really is useful and necessary to use a genericized term in your case.

11 Comments

  1. There is an early scientific article dissecting the term VGI by Francis Harvey, now Professor at the Institut für Länderkunde, Leipzig, Germany :

    Harvey, Francis (2013): To Volunteer or to Contribute Locational Information? Towards Truth in Labeling for Crowdsourced Geographic Information. In: Daniel Sui, Sarah Elwood and Michael F. Goodchild (eds.): Crowdsourcing geographic knowledge. Volunteered geographic information (VGI) in theory and practice. Dordrecht, New York: Springer, pp. 31–42.

  2. Pingback: weeklyOSM 473 | weekly – semanario – hebdo – 週刊 – týdeník – Wochennotiz – 주간

  3. I think the first one who used the term was Goodchild (Goodchild, M. F. (2007). Citizens as sensors: the world of volunteered geography. GeoJournal, 69(4), 211–221.).

    You have to keep it mind that he used it in a particular context and he used it to describe (or maybe more trying to wake up some geographers) “the widespread engagement of large numbers of private citizens, often with little in the way of formal qualifications, in the creation of geographic information.”

  4. Pingback: [Перевод] Новости из мира OpenStreetMap № 473 (06.08.2019-12.08.2019) | Компьюлента

  5. Pingback: [Перевод] Новости из мира OpenStreetMap № 473 (06.08.2019-12.08.2019) - Badwebs.ru

  6. I have never meet Goodchild, it will probably be good to ask him his point of view on that.

    But as i say I think you should take the context of the use of this term in 2007. I was quite new on the geographic field in 2007 (and I still feel green) but if I remind correctly lot of geographers (in the academic) didn’t see how smartphones, web and geolocalization will have effects on interacting with the relationship everyone have with space/place. The term “VGI” helped think something new at this moment.

    Was it a good help ? well I don’t know. Should a better term be use ? I guess probably especially with “corporate editing”, “machine learning mapping” and not so “volunteere” on collecting data.

    Can I ask you what make you feel it is overused or that it produce over generalization ?

    • As said i do not primarily criticize the use in 2007 – for all i know the term was just used to make a point back then and it does not really matter much how valid that point is from today’s perspective.

      I also don’t criticize the volume of use today, i criticize that AFAIK nowhere in the scientific literature that uses this term more recently there is a critical reflection on its suitability to characterize the subject it is used for. This leads me to the conclusion (and this is a realization i came to after i wrote this post in 2015 by the way) that the use of this term is overwhelmingly not due to a desire of precisely characterizing the subject of their research but due to pressure to use certain buzzwords to be taken seriously by their peers (and likely by funding agencies).

      I brought up this subject again now (which was echoed by weeklyosm and others to a larger audience) to raise awareness among people in the scientific domain who might be non-consciously using this term about this problem and about the fact that outside the peer group in their specific scientific domain they might actually be making a fool of themselves by thoughtlessly throwing around buzzwords like this.

      If you are doing research on OpenStreetMap or other crowd sourced geodata using the term ‘VGI’ without critical reflection will likely weaken your scientific message and credibility rather than supporting it.

      • Well thanks for the reminder ! In my surroundings (French) the lack of criticisms is not really a problem to say the least so it is a surprise but we tend to publish in french. I can even make a point that we are spending way to much time on critics and not enough time producing concept that can help (especially on everything “digital” related).

        I don’t think the subject is well delimited and it is also a moving target. I think people are moving now more toward “citizen sciences” as a buzzword (and probably because a huge amount of “VGI” is not accessible).

        I will definitely use VGI in an abstract or a speech specially if the public is not from the domain so they start to have an idea. But for the analysis part or conceptual part I will use Desrosières (https://www.hup.harvard.edu/catalog.php?isbn=9780674009691) work or Didier’s one (https://smhs.gwu.edu/epidapo/faculty/emmanuel-didier).

  7. So I teach a lot on VGI and OSM, and I tell my students that OSM is a prime example of VGI. Why? Because it is definitely volunteered (someone made an active decision to add something to the database, Christoph writes that himself in the third bullet point), it is geographic (about data points with a geographic location), and it is also information (because assigning tags etc. is an act involving semantics and other sources of prior information, so it is an interpretation, a cognitive process, more than simple perception and recording of a stimulus).
    I disagree with most of Christoph’s arguments from the blog post. First, using terms like VGI is not an attempt at generalization (removing details), but at categorization (grouping similar items), to facilitate meaningful discussion. Unfortunately, people (including researchers) use terms indiscriminately and inflationary to push their work so that it turns up in searches and gets cited more often. That is the only part where I agree with Christoph, that often buzzwords are used whether they fit or not, because they are, well, buzzwords. We could do away with categories and discuss every project idiographically, but since (natural) science is about finding similarities and differences, what’s the point? Second, on who can be a volunteer, this is one of the few things where literature seems to agree, so I don’t understand why this post introduces artificial ambiguity here. “Voluntarism” is to my knowledge always associated with individual citizens (who can organize in groups or NGOs of course). I have never heard of a satellite image provider (or owner…) to be termed “volunteer”. I know that there are bulk uploads from institutional sources, but that was never the intention, nor is it the majority. The next point on “information” I do not understand. Of course OSM is made up of data, but OSM (a map!) is full of semantics, so calling it information definitely makes more sense. Lastly, why associate or equal volunteered information with private information? If something is out there in the world but not on OSM, then of course I can volunteer to bring it to OSM, and it is verifiable by others. VGI is not about things only I can know. I really do not understand where this notion of information as property (in the context of VGI) comes from. I have not seen it being discussed in that context. Lastly, “crowdsourced” implies a hierarchy: There is someone (a person, an organization, a company) who needs something done but doesn’t have the resources. So they outsource the task to a crowd. Does that describe OSM? I don’t think so. To me it seems the exact opposite of what Christoph tries to argue previously.
    So in summary, I think that Christoph has a point in that too many people use terms like VGI without proper thought or to optimize citation metrics. But OSM not being volunteered geographic information? I disagree.

    • Glad to see arguments directly countering my analysis.

      I am going to answer point by point and hope this will help better understanding my arguments and my critique.

      […] First, using terms like VGI is not an attempt at
      generalization (removing details), but at categorization (grouping
      similar items), to facilitate meaningful discussion.

      I would argue that any categorization is a form of generalization. But it does not matter how you call it – by putting OSM in a category with a lot of of other projects with very different aims, paradigms and organization methods without reflecting on the differences you are not necessarily making a wrong statement (as long as there is some commonality between them in some aspect) but you also do not necessarily make a statement that contributes to a meaningful discussion.

      […] We could do away with categories
      and discuss every project idiographically, but since (natural)
      science is about finding similarities and differences, what’s the
      point?

      Here might lie a source of our differences in view – i consider most of the reasearch that is being labeled to be about VGI to be social sciences rather than natural sciences.

      Second, on who can be a volunteer, this is one of the few
      things where literature seems to agree, so I don’t understand why
      this post introduces artificial ambiguity here. “Voluntarism” is to
      my knowledge always associated with individual citizens (who can
      organize in groups or NGOs of course). I have never heard of a
      satellite image provider (or owner…) to be termed “volunteer”.

      Ok, here we need to carefully look at the terminology. There is a significant semantic difference between the concept of “the volunteer” and the concept of “to volunteer something”. A volunteer is essentially someone who volunteers his/her work time (and only their work time) to a certain purpose.

      If this helps – in German i would probably translate these terms in this context as “Freiwillige(r)” for volunteer and “etwas zur Verfügung stellen” for “to volunteer something”.

      Goodchild’s motive to use the term “to volunteer something” probably stems from his focus on the benefit of individual local knowledge (which is directly contributed by those who have it) which puts crowd sourced mapping at a significant advantage over the authoritive mapping done by official institutions which can only indirectly source this local knowledge in a very work intensive way. But unfortunately this consideration is not part of the term chosen and as explained this is usually lost when people now use the term and cite Goodchild.

      The next point on
      “information” I do not understand. Of course OSM is made up of data,
      but OSM (a map!) is full of semantics, so calling it information
      definitely makes more sense. Lastly, why associate or equal
      volunteered information with private information? If something is out
      there in the world but not on OSM, then of course I can volunteer to
      bring it to OSM, and it is verifiable by others. VGI is not about
      things only I can know. I really do not understand where this notion
      of information as property (in the context of VGI) comes from. I have
      not seen it being discussed in that context.

      These two points we need to consider together. My understanding of the terms: Information is the semantics of data detached from the concrete data representation and independent of its perception by humans. Like data information does not in any way have to be “correct”.

      Knowledge is information available in the human mind that has proven to be intersubjectively verifiable and thereby is considered to be correct.

      What OSM aims to collect is local knowledge. The mapper contributes this in the form of concrete data entered into the database. The information level in between those two is not really of interest – we are not interested in information that is not knowledge (i.e. that is not verifiable or that has not yet made it into the minds of human beings) and neither the mappers nor the project itself claims ownership of the information contained in the concrete data in its database. Therefore while the statement that OSM is volunteered geographic information is not necessarily wrong it is not really a meaningful statement regarding what OSM is about.

      That data can be property under certain circumstances has been established – most visibly in form of EU database law. That knowledge cannot be property in a society with fundamental freedoms is quite self evident – if information can be independently verified it can also be independently gathered (patents, which grant exclusivity on economic use of certain knowledge but not control of the knowledge itself notwithstanding). But the situation regarding information in general that is not verifiable is not quite as clear.

      Official mapping authorities which Goodchild had in mind as a counterpoint to crowd sourced data projects traditionally have a fairly broad interpretation of their ownership of cartographic information. Much of this derives from traditions from pre-democratic times where these authorities indeed had de facto control over much of the cartographic information in their maps and also made use of that by deliberately falsifying information where deemed necessary by the authorities. There are many countries where collection of geographic information is still legally restricted (and even in Europe it largely was until the end of the cold war). This idea that information (and not just its concrete data representation) can be property is still quite widespread today – in particular in the domain of cartography but also elsewhere, just think of the recent discussion of the EU copyright directive which also tries to push such ideas.

      As said i am not sure if this played a role in coining this term “VGI” originally. I find it remarkable and significant that despite Goodchild’s emphasis on local knowledge of people as a major advantage of crowd sourced geodata production he chose to use the term information. Independent of what Goodchild thought i think the success of the idea of VGI might in particular also have happened due to it resonating with people who like the idea that information (and not just data) can be owned.

      Lastly, “crowdsourced”
      implies a hierarchy: There is someone (a person, an organization, a
      company) who needs something done but doesn’t have the resources. So
      they outsource the task to a crowd. Does that describe OSM? I don’t
      think so. To me it seems the exact opposite of what Christoph tries
      to argue previously.

      Although i’d disagree that “crowd sourced” implies someone controlling and managing the crowd as a term itself it has been used mostly for crowds that are steered centrally so for reasons of historic practice it is probably not the best choice to avoid misconceptions.

      Of course OTOH OSM is not in any way immune to attempts at steering from the outside so putting it in a category with other projects where such steering routinely happens is not that far fetched.

      So in summary, I think that Christoph has a
      point in that too many people use terms like VGI without proper
      thought or to optimize citation metrics. But OSM not being
      volunteered geographic information? I disagree.

      Note i never said OSM is not VGI – i explained why i think VGI is not a useful term to subsume OSM and other things under and why using this terms when writing about scientific research is usually a bad choice.

      I hope these comments helped better understand my reasoning – your arguments helped me to better understand how you can see this fundamentally differently, thanks for sharing this.

Leave a Reply

Required fields are marked *.



By submitting your comment you agree to the privacy policy and agree to the information you provide (except for the email address) to be published on this blog.