If agency matters (and I think it does) then maybe it's not such a good idea to reduce people to their sensing/recording/forwarding capability. "Sensor" has a very mechanistic connotation even when we talk about bio sensors. People on the other hand often make informed decisions as part of data collection activities.

Can www made by Caesar in 44 bC be the first social network¿

.

It might not be exactly a valueable contribution, but is it even important to discuss about the sense of that term? After all, it is, more or less a buzzing word describing the collection and emission of data related to an individual by small personal devices. In that sense it is much shorter than the previous sentence and has a certain ring to it and might be used in the same way as "the internet of things" or similar terminology, i.e. not if you want to be precise, but without too much hesitation in abstracts, introductions and when you talk to the press or the general public.

@ Sven Schade

People as sensors, a good idea, but with limitations or boundaries. GIS generally relates to spatially referenced data. Human perception of space is mostly erroneous, even information stored and analysed in mind and decision made through perception over years (cognition) may be considered only as *opinion* not a true assessment of location, spatial dimensions or direction that are to be rectified either through measurements or through an adequate number of testimonies i.e. opinions of different people. Same applies to attributes which is influenced by a sensor i.e., person's previous experiences and bend of mind. For example, if I visit a slum, may be I find it a damp and stinking place, a more sophisticated person may describe it in even stronger terms, while a person used to these circumstance may not be so critical. Angle of vision is also important, from one angle it may give an impression of a specific geometric figure, seen from a different angle or place, it may turn out of a different geometric figure. I am a bit sceptical of VGI project. Its success or failure would provide to review it.

I think 'people as sensors' requires that acquired data or information is gained by human observations, not by only using traditional physical sensors integrated in devices like mobile phones.

The measured parameters, which contains most important aspects of a observation have to be sensed by humans. People are much more flexible about different kind of parameters than traditional physical sensors.

To state observed phenomena or events more precisely by using sensors which everybody is carrying around is still improved "user generated content" (e.g. a report of damaged buildings after an earthquake with supporting GPS-Coordinates) and not a measurement of fixed number of predefined quantifiable parameters (e.g. temperature, humidity or earthquakes magnitude).

So, I think the deciding factor is, whether core-data acquired by using humans senses which can be supported by additional sensors or measured by a physical sensor and only transmitted by a human.

I'm absolutly ok with that term, even in german it's a bit tricky to distinguish two meanings:

1. Active: People contributing by measuring or analysing things

2. Passive: Like a mechanical just be part of a toolchain and only be a part of data gathering

As to my experience in UGC / Open projects, most people feel good just by collecting stuff. So even if they are full members, they contribute to the database but think on how to do further development on the project (infrastructure, processes, ...).

People are sensors: they collect and publish data. The real contemporary challenge is to disseminate the huge amount of data into a valuable information in a geographic context.

Hi Sven,

your question is very well poned. Recently i'm following a smart-city research where people may be used as sensor to collect noise data or information about architectural barriers.

I think that the boundary between simple mash-up and people as a sensor is not well identified. If we can consider people as a sensor we should associate to these people/sensors some characteristics typical of sensors, for instance in a sensor we are able to know exactly his position. Is it true in case of people? Yes with some known uncertain if we consider GPS positions in mobile phones, however in several case we don't have this data.

Second, the sensors are programmed to record information in the same way if not accurate information are given to the user to collect data, probably the information will not be usable.

Anyway I think that this volountary based information is a great opportunity, but we don't make the error to confuse mash-up and crowd-sources with people as sensor approach.

If i am not wrong,one of best example of VGI is Google earth. There there is only Images with minimum of annotation. People who are viewing the portion of the system can update the information Like City name,popular places, Street names etc.,But only disadvantage of this system,people may commit mistakes, This leads to the problems. It is good that people can be used as sensor (or update the information) but not spatial related information

If crowdsourcing is considered in GIS, then people should be considered as sensors. The only issue is the accuracy of the data captured. Subjecting the data to much filtering might give the data a reasonable accuracy which may be needed for any average project when no better source is available. Anything is sometimes better than nothing, so having the people as sensors where is no better sensor is alright.

Wow, quite some comment to digest... for the moment, just on terminology:

Internet of Things (IoT) is certainly something broader and on a lower level, i.e. not carrying the specificities of location, geospatial information or even environmental information. Then we should distinguish between people carrying some device, collecting data with that and publishing that data. In such a case, a person is a kind of sensor platform, while cameras, microphones etc. are the real (physical) sensors. Already if people upload data together with some keywords, they begin to give additional information; from here on we might talk about a ‘people as sensors’ METHAPHOR., i.e. stimulus in, internal processing, digital data out. However, one issue remains: people interpret when they perceive. In other words, can we use the sensor metaphor although humans do not almost deterministically convert what they see? Having said this, I start to wonder if already the decision to upload something matters… does this go too far?

Also interesting that nobody (until now) said: ‘calling me a sensor is discriminating’. I have heard this before... and some people shifted to call this ‘citizen surveyors’ ;-)

Sven thanks for a great question I have been contemplating for some time with my fellow collaborators. I have found the contributions of everyone on this list so far to be very stimulating. Here is a TEDx talk on the possibilities of embedded sensors of all types in humans that I gave four months ago or so. The talk was called "microchipping people" but please note I am studying the potential for integration across a variety of levels of infrastructure. Sensors on people will only do clever things if we have the infrastructure to support them-- so consider my attempt to paint a picture of three futuristic scenarios to be bringing together a variety of network types, device types (luggable, wearable and implantable) for ID, location, condition monitoring of the person. Here is a paper in The Proceedings that might also be relevant to the discussion: Planetary-scale RFID Services in an Age of Uberveillance http://works.bepress.com/kmichael/186/

http://www.youtube.com/watch?v=fnghvVR5Evc

Many interesting comments, time to reflect a bit:

Mohammad made a sharp observation. Indeed with VGI (or user generated content in general) much more people speak about places, simply because this is closer to human thinking. This is a fundamentally different approach to (sometimes more precise) location-based systems. Some people – already for several years – claim that mobile devices, such as smart phones, will be all equipped with GPS and thus only location will count. I have my doubts.

Also Joachim made a good point. In the end user contributed content might be a valuable complement to classical sensing results and data sources from officially recognized authorities. But how can we combine both? Which role will trust and reputation plan in this?

About Matthias’s comment, I like mostly that he starts talking about the process and user motivation to contribute to a project. This is certainly a key in active participatory sensing, where users should contribute to reaching a specific goal or to kind of opinion mining. Providing contributors with feedback, i.e. telling them how they contributed on in which ways their data was used, is already a powerful concept, but more direct rewards, such as community points or real money might be others. Any experiences our there?

I also fully agree with Mohamed. The real challenge is to find out where such information can be useful, depending on the phenomenon under consideration and also on space-time (e.g. people rarely detect an oil spill when it occurs, but could provide valuable inputs once it reaches the coast).

Some of Federico’s points are somewhat covered in the above. But, yes SmartCity projects are quite active in this area, which is great. The concerns about metadata and then derived quality, in the sense of fitness for purpose is certainly valid. This might be the right place to point you to a set of five ‘Citizen Observatory’ projects which will be kicked-off in the coming months, all funded by the European Commission under Framework Program (FP7). This is not self-promotion at all, I am not involved, but I guess Federico already knows this ;-)

@V.: I would not call Google Earth a VGI example, it is rather a prominent platform for publishing such information. There are others such a bing maps. It is worth to take a close look into IPR issues before deciding to publish, e.g. (as far as I know) Google gets the right to re-use everything for whichever purposes.

Apart from the quality issue, which is shared by many of you and I also tried to address above, Edet has a interesting view: people are sensors if the result ends up in a GIS. Why not if I publish on Flickr, Youtube or others?

Thanks also to Katina, not at least for the links. I will have closer looks soonest and am quite exited about this stuff. Although some of the previous arguments are repeated you particularly stress the importance to see any (user contributed) data in context, even on the device and general infrastructural level. I believe that this is an important point. With all this data tsunami coming up (and VGI very much contributes to this), context analysis and data mining, simply speaking statistics will become more important than ever… if anybody is in this position: I would immediately enroll for a Master Program that covers this field ;-)

So far for the moment, hope this stimulates further thoughts… have a nice week end!

People as sensors is, indeed an interesting concept. There's a project ongoing (and soon to be anounced to the world at the NOAA/NSSL where people will be asked to submit precipitation reports using iPhone and Android apps. These apps use the GPS to provide precise geolocation, but both the report and the location are needed.

One thing to consider when receiving imprecise locations: Ambiguity. Since all cell phones now have, or soon will have, GPS, we assume their location is well known and relatively precise. However, they all use just a code-phase determination in 3D. In a perfect world, we should expect to be within about 6-10 meters horizontally and 8-14 m vertically (ellipsoidal height) but there are limitations promoting ambiguity in that measurement. Thus, on a marginally designed GPS antenna with minimal receiving engine (long ago, cellphones eclipsed the original processing engines for hand-held GPS, so that's not an issue), issues associated with lost satellites (and thus, poor constellation selection and reduction of overdetermination) and multipath must be considered in an error budget. Similarly, reduced sky visibility when the phone is in a pocket or holster, or is indoors, must be considered. When GPS signal is lost, some phones retain the last known position; others give a no-location response.

Note that I'm discounting problems with the human sensor but highlighting potential issues with precise geolocation. Why? Because the uncertainty of location should be a prime concept when you try to figure out where these reports come from. Allowing a larger error ellipse is usually a good prospect when you have concerns about both precision and accuracy.

@Katina: Great talk, equally mindful, entertaining and frightening. Issues around intellectual property, privacy and ethics will sooner or later change drastically; it is just a matter of time. The risk of ‘society becoming technology’ is certainly worth to be aware of, no matter if peoples are sensors or being sensed for various good and bad reasons.

I agree if crowdsourcing is considered in GIS, then people should be considered as "inteligent" sensors. My past doctoral student Miroslav Kubasek introduced crownsourcig application to support environment protectiion, see www.ZmapujTo.cz. You can translate from Czech paper http://www.priroda.cz/clanky.php?detail=2035 more information. The question where are the boundaries of ‘people as sensors’ opens futher topics in eParticipation and eDemocracy. Above application with voluntary approaches of people brought new impuls to public for environment protection.

If agency matters (and I think it does) then maybe it's not such a good idea to reduce people to their sensing/recording/forwarding capability. "Sensor" has a very mechanistic connotation even when we talk about bio sensors. People on the other hand often make informed decisions as part of data collection activities.

@Sven Schade

I like you, I took ‘people as sensors’ as a METAPHOR otherwise with equipment people become platforms, not sensors. To me to be sensor is not objectionable as we do it every moment. If we cross a road our decision is based on our senses.

My above response is based on distinction between crowdsourcing and VGI, the idea as defined by Michael Goodchild in 2007 centres on humans as sensors that voluntarily create, assemble and disseminate geographic data. This definition is directly in line with crowdsourcing, but it seems to have taken a bit of a turn for some.

While the sharing element remains intact, the alignment with Esri’s Community Maps Program relates the concept more toward professional geospatial content providers rather than the groups creating crowdsourced maps. The Community Maps Program aims to harness the volunteered geographic information into one common base map that can be used by anyone. The fusion of data sources via the Esri-hosted service provides a means for data creators to share their information while still retaining their intellectual property, and some control on how the content is used. Municipalities, states and other entities have opened up their data in this fashion, given the measures for control. However, this idea of VGI nowhere mentions use of GPS or cell phones etc. It speaks of information generation as by humans as sensors, naturally as humans we use over senses to create information and knowledge. However, in the present technological era we may enhance our senses using certain equipment. But, original definition of VGI does not speak of use of equipment.

Hi Jiří, yes, good extension of the terms. Maybe we should say 'people are smart sensors', I got a similar comment on twitter (@innovatearth). Also the widening of the topic to eGovernment is very interesting, this basically feeds into discussions of opinion mining and transparency. Thanks for the pointer.

Christopher's comment addresses exactly what I wanted to mention in my resume yesterday: METAPHORICALLY the notion might fit well, but strictly speaking each input to human sensors (here I mean the sensing organs) undergoes a quite complex process and is always influenced by experiences, cultural orientation etc.

The first part of Mohammad's new comment of goes in a somewhat similar direction. Yet, we might distinguish what humans perceive via their sensing organs and what they then provide as output on a Web 2.0 platform such as Facebook, YouTube, Twitter etc. Two years ago, we discussed another application of the sensing concept: instead of sensing phenomena in the real world, we suggested to monitor VGI platforms for spatial-temporal patterns in the stream of up-loads. This establishes a kind of (virtual) VGI sensor. For example, you could detect that suddenly many people in a short amount of time and located in a particular region start talking about flooding or a forest fire. This could help to attract attention to related events and trigger further action. What do you think about that?

And a sharp observation again 'original definition of VGI does not speak of use of equipment'... still definitions may change, but we should always be clear about which one we are applying. Crowd-sourcing is a brought and very interesting notion, too, when this even goes to applying for obtaining funding from individual citizen we would get off topic... should we explore this a bit?

@Sven: "Also interesting that nobody (until now) said: ‘calling me a sensor is discriminating’. I have heard this before... and some people shifted to call this ‘citizen surveyors’ ;-)"

Actually, for some time I have been contemplating that classic Star Wars phrase from the bar tender: "We don't serve your kind here!" Skywalker responds: "Huh?" And bar tender adds: "Your droid... we don't serve them here"...

All of this came to mind particularly in June of this year when Professor Steve Mann was approached at a McDonald's store in France for wearing his eyetap device... see here: http://eyetap.blogspot.com.au/2012/07/physical-assault-by-mcdonalds-for.html#!/2012/07/physical-assault-by-mcdonalds-for.html There are a multiplicity of complex things that took place here... I leave it to your interpretation...

Steve also contributed a paper to IEEE Technology and Society Magazine on this theme that was published in the Fall 2012 issue. He talks about discrimination... his rights... sousveillance is a very important aspect of this power vs autonomy dichotomy.

We have since decided to host an international symposium on this theme in Toronto next year in June 2013 with the title Living in a SmartWorld and sub-title: The social implications of wearable computing and augmediated reality in everyday life. More on that when the site goes live in a few days.

As humans already are 'electrically charged'- you could say we are sensors today... but my husband (and collaborator) and I have considered the potenial of humans becoming bearers of technology of sensors/nano-devices and the like... at that point are we still human or something else? We have tested the idea of calling this "sensor-based person" an Electrophorus coming from the Greek words (amber/light and to put on). We could thus speak of an evolution taking place beyond that of the Homo Sapien... a person who invites technology into the body could belong to the species of Homo Electricus... someone triggering a device embedded in the body would be interacting with Mr Homo Electromagneticus...

Some more papers offering food for thought: http://works.bepress.com/kmichael/255/ another article in The Proceedings capturing the last 30 years of social implications of technology research... we address some of the themes discussed on this Q&A so you might be interested to read it. And an older article on Homo Electricus. http://works.bepress.com/kmichael/16/

http://4.bp.blogspot.com/-YWAsXMW9HY0/UASnLb34HTI/AAAAAAAAADI/WbWfcf_AAZs/s640/perps123.jpg

@Katina: You are adding two new directions to the discussion.

(1) Providing several points on surveying as ‘that those on top are watching those underneath’, compared to the sense of surveying as applied in geodesy and cartography, which I intended to refer to above. Considering the initial question, I did not think about this point at all, so thanks a lot for adding this facet.

(2) The more and more sensor-mounted person, kind of in line with the notion of ‘smart dust’ which is used in the Sensor Web community for at least the last 7 years. This could be complemented with ideas, such as human-bio-monitoring, in which people do not only wear sensors, but also visit medical doctors in fixed iterations to get several measures done and published. This whole area is kind of a extension and amplification on the human-as-a-sensor- platform concept, as mentioned by several others above.

@Jiří: I just came across an article about ”Power of the People: Crowdsourcing Political Change” might be interesting for you (http://ivn.us/2012/10/21/power-of-the-people-crowdsourcing-political-change/).

Rather than "people as sensors" it would be more appropriate to focus on the activity being performed. "Crowdsourced sensing" or "people based automatic sensing" seems to me more adequate.

@Alberto

Well, this certainly depends on your motivation for using this, you might (at least distinguish):

(1) Using the term for precisely describing a situation; in this case either people reporting on some observation, carrying (physical) sensors, or both. In such a case, 'Crowdsourced sensing' might relate to the first. However, personally, I am not sure if this is what most people are looking for. At least I am more up to:

(2) Using the term to attract attention about the topic and the overall concepts behind. In such a case, you should not use any jargon (such as 'crowdsourcing' ) or be extremely concerned about use of words. The simpler the better. I believe that this discussion would not have gotten over 300 views (basically during a weekend) if I would have asked something like 'what do you think of 'people based automatic sensing'?'

Could you agree to that?

@Sven

Given this motivation, I surely agree!

I believe that "Social sensing" has also been used for the overall concept and pointing to social computing. Moreover, the systemic opposition of the terms "sensing/sensor" with "acting/actor" hints to the technological opportunity to collectively "sense" added to the traditional "social action" concept. A further connection could be made with the "social construction" of the space/reality being sensed.

Even if "social sensing" is a good definition, "people as sensors" is more intriguing. I noticed that what draws attention is precisely the missing verb in "people [...] as sensors".

One thing is considering --as some above do-- that people are *used* as sensors, with a negative connotation: by whom? Are they happy about this? And so on. Another is considering that people are *acting* as sensors, because their sensing is a purposive, shared activity that aims to a common goal, as in sousveillance. Yet another thing is considering people *behaving* as sensors: this point to the fact that interacting with a stimulating environment while being connected may generate data that eventually may even shape an higher order "nervous" system...

Going back to your original question, don't you think that this need to "fill the dots" is precisely what gives sense and opens the boundaries to research?

Excellent comment, also because you apparently are not one of the Sensor Web, VGI, Geospatial Information Science guys, which usually discuss this notion. It is great to get some 'external' views, thanks.

Can you believe, that I am personally very much engaged with the above mentioned communities since five years or so, but never heard about 'social sensing'? Now I already found some papers and will certainly follow it up.

I used to summarize and comment on the discussion at the end of each day, but reading through your statement around the dots, I believe that I could not do any better.

@all: social sensing like "social living" has been around a while... "people as sensors" has also previously been described... look at the patents of Professor Steve Mann... the work by Microsoft Research Labs here: http://research.microsoft.com/en-us/groups/cml/ and the work of MIT Media Labs (e.g. Being Digital is a 1994 book by the then director Nicholas Negroponte) etc... I certainly appreciate the discussion however, esp by Geospatial guys whom I think are in a box position to discuss this topic more broadly!

Sven, if you are interested in this topic maybe you want to talk to Dr. Bernd Resch at the University of Heidelberg. He is deeply into people-as-sensors and his approach "Live Geography"

http://www.berndresch.com/

Yea...I know him since more then 5 years ;-)

^^ "die welt ist ein dorf"

Dear all, thanks a lot for the many inputs and stimulating ideas. Based on the directions you indicated above, I just re-launched the same question (with a slightly modified description) related to five other topics: Internet of Things, Ethics, Social Media, Web 3.0, E-Government… let’s see what these folks are up to. For the moment I did not provide them with a cross-reference in order to avoid bias.

You can follow the new thread here: https://www.researchgate.net/post/Should_we_really_talk_about_people_as_sensors?ev=prf_cont_person_post_question_object

Please feel still free to add further comments!

Tal vez el termino correcto seria personas que son detectadas por el uso de algun tipo de sensor.

@Hugo: yes, for some cases this is certainly an option. If considering the identification of people by sensors, I would still distinguish (1) the real identification of an individual revealing his/her identity, e.g. by face detection software, linking to a mobile phone or some unique characteristic behaviour; (2) the identification of an individual without revealing identity; and maybe (3) tracing of the movements of any of the two.

Here a quick summary of the 'other' discussion thread, i.e. a similar question but flagged with other topics (started 12 days ago):

- Compared to this very active thread, the other just got 66 views and 3 answers so far (one of myself).

- The expected ethical concerns were not raised.

- The distinction we are proposed here seems all right.

- Two interesting examples were mentioned by Fernando Tricas: (1) "google flu trends where the searches of people are used for detecting incidence of flu"; and (2) "Tweet Earthquake Dispatch http://earthquake.usgs.gov/earthquakes/ted/ where the tweets of people can be used to detect actual earthquakes".

I am sorry I do not agree with respected contributor, Christopher Lueg. If we go through history and philosophy of science, science could not reach the present level of sophistication, if people have denied themselves as sensors, using their sense organs to collect information/make observation to reach at scientific laws and explanation of patterns and trends in occurrences (theory building). Humans have always functioned as sensors not only for the sake of progress of science but for their survival. In latter period of our history we enhanced our perception or reach of sensory organs through instrumentation. I do not find fault with the use of word sensor with humans because it sounds very mechanistic. Whether mechanistic or not we have always been sensors and if this capability of us is used for a noble cause what is derogatory in it?

Yes, there is no doubt about people using their senses, particularly because the will to survive. Certainly this also enables science in the first place. However, - and here I assume I am with Christopher - there has been some restrictive way of using the 'sensor' metaphor for describing humans and the way they might provide input to earth observation systems. This is particularly restrictive, because humans do not only perceive some stimuli with their sensing organs, but also process the diverse inputs (vision, sound, smell, touch...) based on their past experiences etc. So the output they are creating depends on the individual person/sensor.

In the end, you might be both right. People are sensing their environment and communicate about the observations. Still, not in all cases can a person be modeled in the same way as a physical sensor, say a thermometer or a camera. As with any modeling, this strongly depends on the purpose of the model. If you want to integrate volunteered geographic information (VGI) - from flickr, Twitter and others - with some authoritative information for detecting some form of event, e.g. a forest fire or a flood, the sensor approach might work well. If you think about behavioral sciences, you might want to look into other alternatives.

That is right. I have already share this thought in my earlier response. Human processing of perception is not as that sensing by some equipment. However, we make observations (senses) every moment either deliberately or accidentally.

Exactly. Good to have this sorted! I very much appreciate your active participation in this! I would also like to take this opportunity to thank all of the contributors, great outcomes so far.

@All: an international conference that you might be interested in, related to the theme of people as sensors. Steve Mann and I are hosting IEEE ISTAS13 in Canada on the theme of wearable computing and augmediated reality in every day life. See link below if you are interested.

http://veillance.me

Brilliant theme and topic list! In fact everything, us and our environment, gets 'super smart' these days, in the sense of being equipped with diverse means to information collection, processing and context dependent notifications. One big question arises: 'Is this for good or for bad - for individuals and our society at large?'.

The scope of this event fits amazingly well to several items we are discussing these days. I will certainly recommend this event to some friends and colleagues, too.

@Sven: Hi. Thanx! It promises to be a great event with a lot of participation. The two pager PDF can also be found below. Feel free to distribute, propose panels etc. We got approval to host this event only recently through IEEE SSIT but we have been planning it since June! Please consider writing or joining with a Canadian colleague who can attend to present the research locally and share outcomes from the event.

@Katina: thanks back. I would love to write about something like 'Quantified Self' (Big Data + Mobile + Sensors + Visualization + Gamification = Quantified Self - http://www.topcoder.com/blog/big-data-mobile-sensors-visualization-gamification-quantified-self/), a notion which I just discovered recently. Let's see if this will work out.