Participatory Infrastructure Monitoring: Design Factors and

Limitations of Accountability Technologies

Dietmar OffenhuberDissertation Defense

MIT Department of Urban Studies and PlanningDec. 11, 2013

Dissertation Committee:Professor Carlo Ratti (chair)Professor Eran Ben-Joseph

Professor Brent D. RyanProfessor Lawrence Vale

Overview Introduction Theories & previous research Overview of the three essays

Essay 1 – Putting matter in place

Essay 2 – Tacit arrangements

Essay 3 – Infrastructure legibility Conclusion

Hand-sorting the refuse of New York City, 1903.From: Melosi, Martin V. 2004. Garbage In The Cities: Refuse Reform and the Environment. University of Pittsburgh Press

Modern Material Recovery Facility (MRF)

Tarzan vs. IBMAlternative title of Alphaville (1965) Jean-Luc Godard

Infrastructure monitoring from the top-down

4IBM Smarter Planet Initiative www.ibm.com/smarterplanet

Intr

oduc

tion

http://intellistreets.com Intr

oduc

tion

Intellistreet fixturehttp://intellistreets.com In

trod

uctio

n

Intr

oduc

tion

Pieter Franken, Safecast - radiation monitoring, Japan 2011www.safecast.org

The Rise of the ‘Expert Amateur’ (Kuznetsov and Paulos, 2010)

Infrastructure monitoring from the bottom-up

7

8

Related work

Citizen participation in infrastructure operationCitizen Co-Management of Infrastructure Schubeler 1996; Ibem 2009

Participatory Monitoring Estrella and Gaventa 1998

Self-organized infrastructure Egyedi and Mehos 2012

Volunteer-driven data collection

Volunteered Geographic Information (VGI) Goodchild 2007; Coleman, Sabone, and Nkhwanana 2010

Participatory Sensing Estrin 2006

Crisis Mapping Okolloh 2009; Starbird and Palen 2011; Meier and Leaning 2009

Social Accountability Initiatives Joshi 2010; Verplanke 2010

Intr

oduc

tion

Participatory Infrastructure Monitoring

Historical changes l in the operation of urban infrastructure services -

– loss of “integrated ideal” of public provision (Graham & Marvin 2001)– blurring the line between provider and user

For better or worse, the individual is more consciously involved in infrastructure systems, infrastructure has become less “invisible” Infrastructural inversion (Bowker 1994)

Intr

oduc

tion

Motivation & overarching questions

Digital augmentation of urban infrastructures and services opens new interaction channels, facilitates self-organization and new forms of participation

What are the implications for the future of urban services and their governance? How do technological interfaces and their design shape these interactions?

Intr

oduc

tion

Conceptualizations of infrastructure

System perspective• Large Technological System (LTS), (Hughes 1987)• Infrastructure as a financial asset (Torrance 2009;

Prud’homme 2005)

User Perspective• Anthropology of infrastructure (Star & Ruhleder 1996)• Inverse infrastructures (Egyedi & Mehos 2012)

Theo

ry

HCI design theory & STS• Conceptual models &

affordances (Norman 1986)• Seamless vs. seamful design

(Chalmers & Galani 2004) • Social presence (Short,

Williams, & Christie 1976; Erickson & Kellogg 2000)

Artifacts and governance (STS)• Software-sorted geographies

(Graham 2005)• Politics of Artifacts (Winner

1980)Conceptual Models: Design Model, User’s Model and System Image (Norman 1986)

Theo

ry

Infrastructure legibility and governance

Legibility from above

“I began to see legibility as a central problem in statecraft [...] the legibility of a society provides the capacity for large-scale social engineering” ---- James C. Scott, 1999

Legibility from below

“the degree to which the inhabitants of a settlement are able to communicate accurately to each other via its symbolic physical features”

---- Kevin Lynch 1984

Theo

ry

Accountability technologiesSocial Accountability: community-driven approaches for keeping power-holders accountable (Joshi and Houtzager 2012; Malena, Forster, and Singh 2004).

Accountability Technologies: the use of social media tools to monitor governance (L. Diamond 2012)

– Coordinating data collection– Collaborative analysis– Activation, insertion into political process

Theo

ry

Overview of the three essays

1. Legibility of structure and activity

2. Legibility of practices and organization

3. Legibility of governance and social presence

16

matrix

Essay 1 – Putting matter in Place Essay 2 – Tacit Arrangements Essay 3 – Infrastructure Legibility

Legibility of system structure and activity of practices and organization of governance and social presence

Locus of the observer outside the system inside the system at the system boundary

Geography Seattle / US São Paulo, Recife / Brazil Boston / US

Unit of analysis Waste Item Worker Reported Issue

Investigated nexus user – providerprovider – provider

provider – cityprovider – provider

user – city

Focus of participation monitoring management governance

Research Questions 1. To what extent can waste transportation diminish the benefits of recycling?

2. To what extent is the final fate of a discarded item predictable through its material and the location where it was discarded?

3. What are the implications and limits of active location sensing for the governance of waste infrastructure?

1. how do waste picker cooperatives and associations respond to data reporting requirements from local governments and companies?

2. How can available location-based technologies improve the data management and coordination of these recycling cooperatives and associations?

1. Which assumptions about the users are embedded the design of online 311 platforms, and how are these assumptions translated into design features?

2. Why do users submit reports, and how do the system’s design features correspond with these motivations?

3. By what mechanisms and design principles do existing systems facilitate and constrain the interaction between citizen and city?

Methods Participatory sensing / quant. Analysis / semi-structured interviews

Participatory Design / semi-structured interviews

Content analysis w. open coding / spatial analysis

Essay 1: Putting matter in place – monitoring waste transportation

Essa

y 1–

Putti

ng M

atter

in P

lace

18

Research Problem

Formal Waste Systems in the US

• Lack of data about waste transportation• Question of transportation is crucial• Calculating the environmental impact of transportation is

difficult• Need for evaluating different collection models

Essa

y 1–

Putti

ng M

atter

in P

lace

19

Research questions

1. Environmental performance: to what extent can waste transportation diminish the benefits of recycling?

2. Predictability: to what extent is the final fate of a discarded item predictable through its material and the location where it was discarded?

3. Policy implications: what are the implications and limits of active location sensing for the governance of waste systems?

Essa

y 1–

Putti

ng M

atter

in P

lace

20

Methods

• Participatory Sensing (Burke et al., 2006), self-reporting location sensors; 96 Households in Seattle Metro

1. OLS Regression with categorical predictors

2. Multinomial Logistic Regression with categorical predictors

3. Semi-structured interviews with waste professionals

• Seattle was chosen because of its aggressive recycling policy and its geographic location (variety of different possible pathways)

Essa

y 1–

Putti

ng M

atter

in P

lace

Trash | Track volunteer with donated objects and sensors (Photo: C. Chung 2009)

Essa

y 1–

Putti

ng M

atter

in P

lace

Essa

y 1–

Putti

ng M

atter

in P

lace

26

Findings RQ1 & RQ2Significantly longer transportation distances for e-waste and HHW compared to curbside collected recyclables (Plastic).

In some cases, transportation neutralized energy savings of recycling (WARM model)

Problematic aspects of mail-back / take-back programs: longest transportation distances, involving air-freight.

Outsourcing of waste transportation to courier services, which are not optimized for the purpose.

As expected, recyclable materials are less likely to end up in a landfill

Significantly higher likelihood for a landfill outcome in suburban / rural municipalities

Essa

y 1–

Putti

ng M

atter

in P

lace

Google maps Screenshot with location reports from waste items at a gravel facility, which is not permitted to accept any kind of waste except construction waste (p.75)

Essa

y 1–

Putti

ng M

atter

in P

lace

Implications for planning (R3)

“Recycling is the password for shipping things to other countries”Obstacle for law enforcement is proving violation and damage under RCRA

Tracking in voluntary certification programs• Audits of sub-contractors – strict liability under RCRA• Incentive for recyclers to join stewardship programs with

mandatory tracking

Essa

y 1–

Putti

ng M

atter

in P

lace

Acknowledgements

Conducting the Trash Track experiment involved a large team, whose contributions were essential for the project: David Lee, Malima Wolf, Jen Dunnam, Angela Wang, Rex Britter, Assaf Biderman, Carlo Ratti and others.

The essay is based on my my first year paper, my own analysis and research design, revised and extended for the dissertation.

My first year paper was also the basis for a co-authored paper published in JAPA 2012Offenhuber, Dietmar, David Lee, Malima I. Wolf, Santi Phithakkitnukoon, Assaf Biderman, and Carlo Ratti. “Putting Matter in Place.” Journal of the American Planning Association 78, no. 2 (2012): 173–196.

Essa

y 1–

Putti

ng M

atter

in P

lace

31

Essay 2: tacit arrangements – data reporting practices of recycling cooperatives in Brazil

COOPAMARE Recycling Cooperative, São PauloEssa

y 2

– Ta

cit A

rran

gem

ents

32

Context and research problem

• August 2010 comprehensive National Policy for Solid Waste (Política Nacional de Resíduos Sólidos – PNRS)1. Extended Producer Responsibility (EPR)2. Inclusion of recycling cooperatives in the valorization

• Tension for cooperatives between new data collection requirements and new opportunities

Essa

y 2

– Ta

cit A

rran

gem

ents

33

Research questions

1. How do waste picker cooperatives and associations respond to data reporting requirements from local governments and companies?

2. How can available location-based technologies improve the data management and coordination of these recycling cooperatives and associations?

Essa

y 2

– Ta

cit A

rran

gem

ents

34

Methods

1. Comparative analysis of cooperatives involving site visits and semi-structured interviews

2. Participatory Design (Kensing and Blomberg 1998)– Mapping workshops with functional prototypes

• São Paulo (Nov. 2011)

• Recife (Jun. 2013)

– Semi-structured Interviews

Essa

y 2

– Ta

cit A

rran

gem

ents

Theory & prior research

Theories of informality – Dualist (ILO 1972)– Structuralist (Portes and Castells 1989)– Legalist (De Soto 1989)– Voluntarist (Maloney 2004)

Formalization Models (Scheinberg 2012)– Service model– Commodity model– Hybrid model– Community-based enterprise (CBE)

Most studies on informality investigates the micro- (individual) or macro-(economic) scale. The meso-scale (i.e. spatial organization) is under-researched.

Essa

y 2

– Ta

cit A

rran

gem

ents

Interview Partners Cooperatives Local Experts

Rede CataSampa João Ruschel Dr. Lucia Helena Xavier (FASPE, Recife)São Paulo, SP Prof. Dr. Maria Cecilia Loschiavo dos Santos (USP)CRUMA Roberto Golfhino Prof. Dr. Flavia Scabin (FGV)Poa, SP Prof. Dr. Tereza Cristina Carvalho (CEMPRE)Coop-Reciclavel Cristiano Mateus Mendoza (Giral) Guarulhos, SP Diogo Vallim (Giral)COOPAMARE Maria Dulcinea João Ruschel (CataSampa)São Paulo, SP Laerte Paz

Manuel Soares Laura FostinoneWalison da Silva Ana BonomiFrancisco da Silva Rafael GalvãoElisonete Ferreira Da Conçeicão

COOCARES Lindaci GonçalvesAbreu e Lima, PE Team MITCOOREPLAST Roberto Antonio Gonçalves Dietmar OffenhuberAbreu e Lima, PE Vânia Maria da Silva David LeePro-Recife Roberta da Santana Pessoa Libby McDonaldRecife, PE Maria dos Prazeres Santana

José CardosoNossa Vida Pedro Lima AlcantaraRecife, PE Adeilda de Lima

André Carlos da Silva CardosoJosé de Iterlao Nefo

City of Olinda Tereza Cristina Angelo da SilvaOlinda, PEEnvironmental Ministry ofthe State of Pernambuco

Ana Gama

Recife, PE

Essa

y 2

– Ta

cit A

rran

gem

ents

Comparison of cooperatives

Essa

y 2

– Ta

cit A

rran

gem

ents

38

Results RQ1

• Wide variation with regards to size (4-80 members), collection models, revenue models and external partnerships.

• Every group reported data the local government on a regular basis. Yet, formalization often stopped half-way, grey area in both law and practice, accepted by both cooperatives and the city.

• Quantifying value of the service – material as donation vs. collection as a service– Understanding true cost of collection– Data considered more valuable for business partnerships than for the

city

Essa

y 2

– Ta

cit A

rran

gem

ents

Recife – Verde é Nossa Vida, Jun 2013 Android application, Jun 2013

São Paulo- Coopamare, Nov 2011 Recife – Pro-recife, Jun 2013Essa

y 2

– Ta

cit A

rran

gem

ents

Recorded collection routes, Coopamare / SP , Nov. 2011

Essa

y 2

– Ta

cit A

rran

gem

ents

Results RQ2

Demonstrated value for (1) coordinating truck collection (Coopamare & Pro-Recife); (2) sourcing e-waste (Pro-Recife)

Issues discovered:– Raising unrealistic expectations from clients (COOPAMARE)– Diminishing face-to-face contact with residents, therefore

opportunities for building trust (COOPAMARE)– Security (Verde é Nossa Vida)– Operation of the device in the work environment

(COOREPLAST)– Illiteracy less hurdle than expected

Essa

y 2

– Ta

cit A

rran

gem

ents

42

Essay 3: Infrastructure legibility – a comparative analysis of open311-based citizen feedback systems

Map of two years of ‘311 calls’ in New York City – relative comparison of noise, litter and graffiti complaints aggregated by block

43

Essa

y 3

– In

fras

truc

ture

Leg

ibili

ty

Context

Essa

y 3

– In

fras

truc

ture

Leg

ibili

ty

45

Research questions

1. Which assumptions about the users are embedded in the design of open311 platforms, and how do these assumptions translate into the design?

2. Why do users submit reports, and how do the system’s design features correspond with these motivations?

3. By what mechanisms and design principles do existing systems facilitate and constrain the interaction between citizen and city?

Essa

y 3

– In

fras

truc

ture

Leg

ibili

ty

46

Methods

Methods1. Comparative analysis of design

factors of 8 different system designs

2. Content analysis with open coding of 2200 submitted reports from a total of 26k submitted reports via smartphone

3. Spatial Analysis of 240k submitted reports via various reporting mechanisms

Location

Boston officially supports two very different systems – CitizensConnect and SeeClickFix, allowing for a comparison of design factors.Both systems have +3 years of data

Essa

y 3

– In

fras

truc

ture

Leg

ibili

ty

Essa

y 3

– In

fras

truc

ture

Leg

ibili

ty

Essa

y 3

– In

fras

truc

ture

Leg

ibili

ty

Findings

1. Two distinct design paradigms - minimal (one-to-one) and social (many-to-many).

2. City initiated systems – emphasis on actionability and service delivery, independent systems – emphasis on discourse and accountability.

3. Reflected in the design features (categories, visibility)4. This has an effect on the discourse: on the tone, on the

kind of arguments reporters make5. Regulatory effect of design decisions - opacity through

transparency

Essa

y 3

– In

fras

truc

ture

Leg

ibili

ty

50

CitizensConnect vs. SeeClickFixEs

say

3 –

Infr

astr

uctu

re L

egib

ility

Opacity through transparency

Essa

y 3

– In

fras

truc

ture

Leg

ibili

ty

Conclusion

• Different places for Participatory Monitoring in governance of waste systems.– Driven by citizens (Essay 1 & 3)– Driven by service providers (Essay 1 & 2)– Driven by cities (Essay 3)

• Design plays a decisive role for shaping interaction (Essay 2 & 3)

• Infrastructure legibility - linking design theories to Urban Planning literature

Future Directions

• Informal governance through mediated interfaces– formalization of informal practices – informalization of formal processes

• “Pattern language” for urban interaction design