Dana Cuff - Professor of Architecture and Urban Design: University of California, Los Angeles Jerry Kang – Professor of Law: University of California,

Dana Cuff - Professor of Architecture and Urban Design: University of California, Los Angeles

Jerry Kang – Professor of Law: University of California, Los AngelesMark Hansen – Associate Professor of Statistics: University of California, Los

Angeles

Communications of the ACM, March 2008/ Vol. 51, No.3

Presented By: Michael Wei

OutlineMotivation for the paperShift in sensing focus and arising concernsStructure and policy of data commonsFuture of the data commons

Shifting From Woods to City• Sensing techniques in the

reserve not scalable to urban setting

• New concerns arise from shift in sensing focus

• Mobile phones are the key sensor platform since it is not economical to deploy fixed sensors in an urban setting.

• New sensing model shift from full centralization to decentralized models and participatory sensing.

James Reserve

Concerns Arising From Urban Sensing

Biased and inaccurate data captured by amateur contributors using uncalibrated cheap equipment.

Incorrect analysis done by unqualified usersSolutions:

1) Distributed system will be able to detect inaccurate data based on multiple data points2) Discussion forums (“social data analysis”) will prune out the incorrect analysis

In other disciplines such as art or politics, data accuracy is not the most important objective.

Beyond Science:D-Tower in the Netherlands

Blue = Happy Red = LoveGreen = Hate

Yellow = Fear

The Data CommonsData Commons – central data repositoryData commons has 4 attributes that

previously defined public sphere (physical realm):1) Must be accessible to diverse members2) Provide opportunity for multiple uses3) Encourage exchange among participants4) Be recognizable as “data commons”

Reasons for Data CommonsPeople express values and participate in

the community through political shopping, eating, and commuting.

To enable these actions, people need access to the data commons to make an informed decision

Subscribing to such participatory sensing is also a form of civic participationi.e. YouTube and presidential debates, placeblog

Building Data Commons

Need to have some way to shape the growth of the data commons

Will require legal, policy, and technological decisions to build a high quality data commons

Legal Implications• Intellectual Property:

- Copyrights only protect creative expressing, not underlying data

- Don’t want to create too stringent intellectual property rights because it might create an “anticommons”

- It is too restricting to have every photo and data piece be copyrighted and have to ask for permission

• Liability:- There is a measure in the Children’s Decency Act

that states that the intermediary parties are not liable for the content their users submit.

Policy - Privacy• Information Privacy - an individual’s claim to

control how personal data is collected, distributed, and processed.

• Self surveillance – people will want the information to be detached from themselves, not traceable.

• Network Solutions: Localization and Time synchronization resolution controls critical to privacy

• Privacy preferences are adaptive. The policy we implement now will have long term path dependent effects

Technology – User InterfaceUser interface determines whether a system

will be utilized or not.Want to be able to easily upload, search, and

share dataEasy data discovery, subscription, and

republication is crucial to data commons

The Fate of the Data CommonsExistence of data commons by itself does not

solve problems since the resource can be misused and under-realized

Predict the future of the data commons by drawing on Greek Mythology

Sirens – Seducing People Into Making Rash Decisions

• Urban sensing will illuminate some characteristics that might otherwise have not been seen such as ground water toxins or toxic waste sites.

• More information does not produce more rational choices. People might give attention to types of data than they really deserve.

• Ex: if data commons offers depictions of violent crime rates in the city, the information might persuade people to move to distant suburbs ignoring the mortality risk that increased highway driving brings.

• Relying on this data might cause a self-fulfilling prophecy where dangerous areas become even more dangerous

• Need to have disclosures or allow comments that deter hasty and misguided analysis

Cyclops• Tragedy of Cyclops – knowing the future and

not being able to change it• Example: Environmental sensors might allow

people to see rapid environmental decline. Those who are lack political or financial clout might see their demise without any way to change it.

• Solution is to “arbitrage our ignorance”: Make predetermined decisions based on ignorance before the data is revealed.

ConclusionSensing has clearly expanded to an urban

setting which offers a new level of community interaction and information sharing. However, we need to make some clear decisions now in order to guide the growth of the data commons.

[1] Cuff, Dana, Mark Hansen, and Jerry Kang. “Urban Sensing: Out of the Woods” Communications of the ACM 51.3 (2008): 24-33.

[2] “Cyclops" Wikipedia, The Free Encyclopedia. 17April 2009, 17:28 UTC. Wikimedia Foundation, Inc. 19 April 2009. <http://en.wikipedia.org/wiki/Siren>[3] “D-Tower” Nox Arcspace.13 Feb. 2006, 19 April 2009<http://www.arcspace.com/architects/nox/d_tower2/d_tower2.html>