Replicate Our Model
Air Partners started in 2017 with the guiding principle of decentering our own objectives related to science and engineering in order to center the priorities of our community partners and their stakeholders. This posture toward our work has led us to a process of continually learning from and with our partners.
Because no two partners have exactly the same priorities, and there is no “one size fits all” solution for stakeholders, all of our projects have taken a unique form. However, in navigating numerous projects in a variety of contexts and with a diverse set of partners, we have developed a generalized approach to community-driven air quality projects that serves as a flexible roadmap for the type of work we do.
We have endeavored to summarize this model in such a way that it can be adapted to a wide variety of contexts, and have included guiding questions at each step that we have found helpful in centering community priorities. As you engage with these materials, please feel free to reach out to us for support. We’d love to learn alongside you.
01 Identify Stakeholder Priorities
02 Plan Project
03 Collect Data
04 Generate Insights
05 Create Artifacts
06 Leverage Artifacts
07 Conclude Project
01
Identify Stakeholder Priorities
Find alignment and create trust
Identify questions that need answering
Key Questions
What are the most important goals for each partner organization? Where is there synergy? Where is there tension?
Which mutual goals relate to air quality?
What are current barriers to achieving those goals? What do you need to learn? What evidence do you need?
What existing information do you have, or where might you look?
What other stakeholders do you need to consider (funders, elected officials, others), and what do they need from this project?
Activities
Understand partner expectation
Establish roles and define team
Understand project endpoints
Establish goals
Site (virtual) visits
Products
Curated list of each stakeholder's priorities
Clearly articulated list of key questions for the project
Stakeholder map(s)
Case Study: Roxbury
-
For the first 3 months of the project, we (Air Partners) engaged in an iterative process of listening and building trust. This took the form of facilitating both formal and informal meetings with coordinators and executors (see Community of Practice diagram) to understand ACE’s priorities as an organization, their activities in support of those priorities, the extent to which those priorities related to air quality, and challenges that they were facing that might be supported by Air Partners. After each meeting we synthesized learnings and brought them back to ACE for feedback and iteration. By the end of this process, we had learned that:
ACE has a high priority of working for the health of children and the elderly.
ACE’s advocacy work related to air quality is primarily related to the impact of transportation and Nubian Station on Roxbury’s air quality.
ACE and Air Partners have complementary strengths (deep knowledge of and connection to groups and organizations in Roxbury, and data collection and analysis, respectively).
ACE has little personnel bandwidth to take on additional work, and thus the “community-led” aspect of the project had to be downsized.
ACE has no in-house capacity to manage or analyze data, and rely on external partners for that work. Our hope of sharing data analysis responsibilities with ACE needed to be reevaluated to include a larger role for Air Partners.
The community of Roxbury is undergoing rapid transformation with significant construction projects throughout the community, with major drivers being gentrification and the expansion and development of Northeastern University.
ACE had successfully advocated for the Massachusetts DEP to place an air monitoring site in Roxbury, but that they have found it to be insufficient due to that site:
Being only one isolated site that offered no spatial resolution in identifying and quantifying pollutant sources,
Generated data that were not easy to access or understand without expertise,
Publicly displayed only air quality index (AQI), which obscures many temporal and pollutant-by-pollutant signatures in air quality.
The ideal air monitoring network would be designed to answer ACE’s questions with monitors near 1) Nubian station, 2) vulnerable populations, and 3) pollution sources (with buses and trains as the priority)
The result of these learnings was a project proposal that centered on three pillars: 1) Establishing an air monitoring network designed to answer ACE’s key air quality questions, 2) Executing a pilot program for deploying HEPA air purifiers in daycares and residences, and quantifying their ability to reduce pollutant exposure, and 3) Generating and packaging insights from data into form factors that ACE can effectively use for their advocacy work.
02
02
Plan Project
Translate stakeholder priorities into a project plan that can answer key questions
Key Questions
What equipment will enable your team to collect the right data?
How much equipment fits into the budget?
How many sites do you need and where should they be?
Who will be responsible for different aspects of the project?
What is your proposed timeline for completing the work (steps 3-7)?
What permissions do you need to begin collecting data?
What supporting materials will help you deploy equipment?
What will you do when things go wrong?
Activities
Write a plan (to include maps, timelines, tech specs, budget, etc...)
Identify and purchase equipment
Develop deployment strategy for equipment (locations, duration)
Provide feedback
Iterate plans based on feedback
Products
Project planning document, including roles, timeline, budget, risk mitigation, and data management
Maps of proposed equipment deployment locations
Case Study: Roxbury
-
During Phase 1 work, our partners at ACE reported being primarily concerned with the impact of transportation emissions from diesel buses, diesel commuter rail trains, and on-road vehicles, in addition to the impact of construction and development. As outlined in the sensor mapping proposal, our priority for sensor placement was to locate sensors where known and suspected sources of pollutants are in close proximity to the most vulnerable populations (older adults and young children). Our overall project budget was constructed to maximize budget for air sensors while providing funding for necessary project personnel and other expenses, and we chose a combination of Modulair (gas + PM) and Modulair-PM (PM only) sensors from QuantAQ based on their unique ability to meet our requirements of accuracy, solar-powered deployment, and remote access to sensor status and data. A more detailed description of our rationale for choosing this number and type of sensors is here.
Our strategy for deploying sensors was to cluster sensors near known and suspected sources of pollutants that represented ACE’s priorities, such as a major bus station, diesel commuter rail line, and major roadways. When choosing exact sensor locations, we sought to place sensors near high densities of vulnerable populations, such as daycares, schools, and elderly housing. We summarized all of this source and population information on a map, and identified 2-4 block radii for ideal sensor locations.
In choosing specific deployment locations, we initially sought permission from the City of Boston to install sensors on light poles for maximum flexibility. This option ended up being bureaucratically impossible, and with ACE we pivoted to seeking partnership with residents and businesses to install sensors. This pivot provided new and unexpected possibilities for community engagement and basebuilding through the sensor deployment process.
03
Collect Data
Collect the right data to answer key questions
Key Questions
Is equipment functioning as anticipated?
How much unanticipated downtime are you observing from equipment? What does this imply about data quality and possible insights.
How and how regularly are data being accessed and organized?
How does preliminary analysis inform any adjustments you need to make in your data collection strategy?
Activities
Deploy instruments in the field
Access, download, organize, backup data
Products
Network of deployed hardware
Organized, backed up raw data that are accessible to the team
Case Study: Roxbury
-
Over the course of 3 months, we worked with partners at ACE to conduct sensor installation trainings and install sensors. While we aimed to have ACE staff and youth organizers participate in all sensor installations, scheduling constraints and availability necessitated the Air Partners team installing many on their own. Once sensors were installed, we monitored their performance through QuantAQ’s online portal and made site visits to address unexpected instrument downtime. Despite the reliability of our chosen instrumentation, sensor maintenance was a task requiring regular attention throughout the project. The two most common issues we encountered were: 1) shading limiting solar radiation at some sites, preventing the battery from charging sufficiently, and 2) a firmware problem requiring a hot-swap of SD cards. Scheduling constraints and sometimes infrequent checks on fleet health resulted in some sensors being down for a week or more at a time, but overall datasets comprised a year or more of complete data.
04
04
Generate Insights
Analyze data to draw conclusions
Key Questions
What figures, tables, statistics, or other data visualizations will help you answer key questions?
What software or algorithms will you use to analyze data and create figures, tables, and statistics?
Who will be responsible for conducting analysis? Do these analysis skills exist on your team, or will you need to collaborate externally?
Do analysis software and algorithms already exist, or will you need to create them?
Activities
Create or obtain existing analysis code
Use analysis code to generate statistics and figures
Interpret statistics and figures
Products
Analysis software and algorithms
List of insights that answer key questions
Curated set of figures, tables, and statistics that support insights
Case Study: Roxbury
-
As datasets were being generated, we created data analysis pipelines in R that pulled data from the QuantAQ database, cleaned data, and generated key figures and statistics. We also developed software to automatically pull data from each sensor in the network and generate monthly reports that are emailed to stakeholders. Finally, we created a webapp that allows partners to easily view real-time data from all the sensors in our network via the Air Partners website. All of this work was completed by undergraduate summer research students and undergraduate students participating in their capstone design experience to fulfill graduation requirements at Olin College of Engineering. We held regular design reviews between ACE, Air Partners, and other stakeholders as relevant as an Agile way to continuously seek feedback and adjust our work to best meet our partners’ needs.
Our pipelines were built using R for two main reasons. First, the Air Partners team had existing competence in R and the Data Science course offered at Olin (which many Air Partners students take before joining the research team) is based in R. Further, and perhaps more important, the openair package in R is a powerful open-source tool designed to facilitate analysis of air quality data like that being generated by this project.
During Phase 4, the Air Partners team conducted multiple codesign sessions with ACE and other stakeholders from Roxbury to identify which figures were most accessible and offered the most utility for their ongoing advocacy work. These codesign sessions also laid the foundation for understanding the specific form factors that were most helpful for artifacts created in Phase 5.
05
Create Artifacts
Display and package insights to tell stories that support stakeholder priorities
Key Questions
For what stakeholders or audiences do you need to generate artifacts?
Do different stakeholders need different artifacts to support their goals?
How will stakeholders use artifacts?
What examples do you have for artifacts that you can use to guide your process of packaging insights?
Who will be responsible for creating artifacts?
What will your process be for soliciting stakeholder feedback and iterating on artifacts?
Activities
Identify key figures and statistics
Incorporate key figures and statistics into formats that community partners can understand and use
Provide feedback
Iterate
Products
Artifacts that have been packaged for stakeholders
Case Study: Roxbury
06
06
Leverage Artifacts
Sharing artifacts to create change
Key Questions
What is your theory of change for how an input of data-informed artifacts create impact?
What is the typical mechanism by which stakeholders engage in advocacy or change-making work?
Who is responsible for coordinating meetings, events, or other mechanisms for disseminating insights?
Who from the team should be present at these meetings, events, or other dissemination opportunities?
Activities
Identify avenues for sharing insights and artifacts
Create a dissemination plan
Implement dissemination plan
Products
Theory of change for how artifacts lead to impact
Plan for using artifacts in change-making work, including roles, timeline, budget, and risk mitigation
Case Study: Roxbury
07
Conclude Project
Develop a plan to either sustain or end project
Key Questions
What assumptions does your team or stakeholders have about the permanence of the work you have completed for the project?
How will you ensure that value continues to be generated by artifacts?
What ongoing support will be needed to sustain key outputs of the project?
Who will be primarily responsible for stewarding artifacts?
How will physical resources be re-deployed or otherwise used for future work?
Activities
Decide the future trajectory of project, physical resources and human resources
Develop and implement project conclusion plan