1. Introduction

While this toolkit has been developed in collaboration with AirGradient, we recognize that AirGradient monitors may not always be the optimal choice for every air quality monitoring project. The principles, methodologies, and community engagement strategies outlined in this toolkit remain useful regardless of the specific hardware you choose to deploy.

Selecting the right monitoring solution

When selecting an air quality monitor, it’s important to weigh several factors to ensure the system aligns with your project’s needs:

• Existing Infrastructure – Your organization may already use certain monitoring platforms or have established data workflows. Switching systems can be costly or disruptive, so compatibility with current tools is worth considering ( see Section 5 on “Limitations and Considerations”)

• Sensor Capabilities – Some projects require specialized sensors (e.g., specific gas detection, higher accuracy, or unique measurement ranges) that may only be available in certain monitoring solutions.

• Budget and Pricing Models – While some monitors are designed to be highly cost-effective, others may offer pricing structures, subscription models, or features that better fit your funding framework.

• Regional Availability – Shipping costs, import restrictions, or limited product availability in your region may influence which solutions are more practical and affordable.

• System Integration – If your project needs to connect with institutional systems, databases, or reporting frameworks, you may need a platform that integrates smoothly with those requirements.

• Research or Project-Specific Needs – Specialized research may demand particular data formats, sampling rates, or sensor configurations that are better met by certain monitoring platforms.

• Long-Term Sustainability – Open-source platforms often provide better long-term support through community contributions, while commercial options may risk discontinuation or pricing changes.

By evaluating these considerations, you can choose a monitoring solution that best supports your technical, financial, and operational goals.

Despite being developed around AirGradient hardware, this toolkit remains highly valuable for projects using alternative monitoring solutions. The foundational elements of successful air quality monitoring projects are universal, and many chapters apply regardless of your chosen hardware. The Planning section, including project definition, community engagement planning, and site selection principles, provides frameworks that work with any monitoring technology. Similarly, the Operations section covering ongoing project monitoring, community engagement strategies, and impact measurement offers methodologies that transcend specific hardware choices.

Key Highlights:

• This toolkit’s planning and community engagement strategies are hardware-agnostic
• Alternative monitors can still benefit from AirGradient’s data platform through OpenAQ integration
• Calibration principles are universal and can be adapted to any sensor system ( see Calibration Alternatives)
• Data analysis and visualization techniques apply to any air quality dataset
• A comprehensive resource directory at the end of this chapter links to key platforms, tools, and communities to help you in your monitoring project

2. Navigating the Toolkit with Alternative Monitors

If you’re using non-AirGradient monitors, we recommend following a modified path through this toolkit that maximizes the value of the universal principles while acknowledging hardware-specific limitations.

For any monitoring project, your first step should be the Planning phase, which is fully applicable across hardware types. Chapter P1 will help you understand the overall approach and methodology of the toolkit, while Chapter P2 guides you through completing the project definition worksheet in a hardware-agnostic manner. Chapter P3’s community engagement planning strategies are universally applicable, and Chapter P4’s site selection principles can be adapted to your chosen hardware with minor modifications for specific mounting requirements or connectivity needs.

The Setup phase requires more adaptation, as Chapters S1 through S7 contain AirGradient-specific setup instructions such as mounting the monitors and connecting them to the internet. These instructions won’t apply directly to your hardware, but you can usually find the equivalent guidance in your monitor manufacturer’s documentation. The calibration principles outlined in Chapter S8 can be adapted to your chosen platform using the general guidance provided in this chapter. The dashboard configuration concepts from Chapter S6 can also inform how you approach configuring whatever data platform you choose to use.

The Operations phase is largely applicable to any monitoring project. Chapter O1’s ongoing project monitoring concepts apply universally, though you’ll need to skip the AirGradient dashboard specifics in Chapter O2 while applying the general dashboard management principles to your platform. The data quality and accuracy maintenance concepts in Chapter O3 can be adapted to your monitoring system, and while the specific maintenance procedures in Chapters O4.1, O4.2, and O4.3 are AirGradient-specific, the general maintenance principles remain relevant.

3. Alternative Data Platforms and Dashboards

When working with non-AirGradient monitors, you’ll need to consider alternative platforms for data management, visualization, and analysis. Several excellent options exist, each with their own strengths and use cases.

Sensor.Community represents one of the most accessible open-source alternatives, operating as a global citizen science initiative that provides an open platform for environmental sensor data. This platform offers access to data from thousands of sensors worldwide, with all data freely available for download in various formats. The platform provides API access for programmatic data retrieval and includes a built-in mapping interface showing sensor locations and current readings. What makes Sensor.Community particularly valuable is its active community of contributors and developers who provide ongoing support and development.

OpenAQ offers another powerful alternative by aggregating air quality data from around the world into a single, open platform. This platform standardizes data formats across different sources, providing global coverage that includes both government monitoring stations and low-cost sensors. OpenAQ is used extensively by researchers, policymakers, and organizations worldwide, lending credibility to projects that utilize its data. The platform offers comprehensive API access for data retrieval and analysis, along with data quality indicators that provide information about data sources and reliability.

For organizations requiring more control over their data infrastructure, the combination of InfluxDB and Grafana provides a powerful, self-hosted solution for air quality data management and visualization. InfluxDB serves as a time-series database optimized specifically for data like air quality measurements, offering high-performance data ingestion and querying capabilities with built-in retention policies for data management. Grafana complements this with professional-grade dashboards and charts, real-time data visualization capabilities, alerting functionality, and an extensive plugin ecosystem.

Key Considerations for Platform Selection:

• Technical expertise available in your organization
• Budget constraints and ongoing operational costs
• Data ownership and privacy requirements
• Integration needs with existing systems
• Community engagement requirements
• Long-term sustainability and platform stability

4. Calibration Alternatives

The calibration principles outlined in Chapter S8 form the foundation for improving sensor accuracy regardless of the specific hardware platform you’re using. These universal principles can be adapted to work with any low-cost sensor system, though the specific implementation details may vary.

The fundamental approach to reference-based calibration remains consistent across different sensor types. You’ll still need to establish a co-location period where your sensors are placed near a reference station (if available) for two to four weeks, during which simultaneous measurements are gathered from both your sensors and the reference station. This data then undergoes statistical analysis, typically linear regression or more complex modeling, to extract calibration parameters in the form of correction factors. These corrections are then validated using independent data before being applied to your network’s measurements.

4.1. US EPA Correction Formula for Plantower Sensors

The U.S. Environmental Protection Agency has developed a specific correction formula for Plantower PMS5003 sensors ( commonly used in PurpleAir and many DIY monitors) based on extensive field testing and comparison with federal reference methods. This formula provides an excellent alternative to local calibration when local reference data is not available:

AGraw <30:
PM2.5 = [0.524 x AGraw] – [0.0862 x RHraw] + 5.75

30≤ AGraw <50:
PM2.5 = [0.786 x (AGraw/20 - 3/2) + 0.524 x (1 - (AGraw/20 - 3/2))] x AGraw – [0.0862 x RHraw] + 5.75

50 ≤ AGraw <210:
PM2.5 = [0.786 x AGraw] – [0.0862 x RHraw] + 5.75

210 ≤ AGraw <260:
PM2.5 = [0.69 x (AGraw/50 – 21/5) + 0.786 x (1 - (AGraw/50 – 21/5))] x AGraw – [0.0862 x RHraw x (1 - (AGraw/50 – 21/5))] + [2.966 x (AGraw/50 –21/5)] + [5.75 x (1 - (AGraw/50 – 21/5))] + [8.84 x (10-4) x AGraw2x (AGraw/50 – 21/5)]

260 ≤ AGraw:
PM2.5 = 2.966 + [0.69 x AGraw] + [8.84 x 10-4 x AGraw2]

4.2. Alternative Calibration Tools

For organizations with technical capabilities, several alternative calibration tools can provide more sophisticated analysis than basic spreadsheet approaches. The R programming language offers several packages useful for sensor calibration, including openair for air quality data analysis, along with standard packages for data manipulation and visualization.

Example R workflow for sensor calibration:

# Load required packages
library(openair)
library(dplyr)
library(ggplot2)

# Load and merge data
sensor_data <- read.csv("sensor_data.csv")
reference_data <- read.csv("reference_data.csv")
combined_data <- merge(sensor_data, reference_data, by="timestamp")

# Perform linear regression
calibration_model <- lm(reference_pm25 ~ sensor_pm25 + humidity, data=combined_data)

# Extract calibration parameters
summary(calibration_model)
slope <- calibration_model$coefficients[2]
intercept <- calibration_model$coefficients[1]
humidity_coeff <- calibration_model$coefficients[3]

# Apply correction
corrected_data <- sensor_data
corrected_data$pm25_corrected <- sensor_data$pm25 * slope + 
                                 sensor_data$humidity * humidity_coeff + intercept

Python provides another powerful platform for sensor calibration, with extensive libraries including pandas for data manipulation, numpy for numerical calculations, and scikit-learn for machine learning approaches to calibration. Python scripts can handle the entire calibration workflow, from loading and merging datasets through performing regression analysis and calculating performance metrics.

For users preferring spreadsheet-based solutions, Excel provides adequate tools for basic calibration tasks. This approach involves importing sensor and reference data into separate sheets, using functions like VLOOKUP or INDEX/MATCH to align timestamps, creating scatter plots to visualize the relationship between sensor and reference measurements, and using Excel’s built-in regression analysis tools to extract calibration parameters.

Key Calibration Principles:

• Co-location period: This should ideally cover the same range of environmental conditions (temperature, humidity, and pollution levels) as during sensor deployment. Usually at least 2-4 weeks for representative conditions
• Statistical validation: Use R², RMSE, and bias metrics to assess performance
• Independent validation: Test corrections on data not used for calibration development
• Documentation: Record all calibration procedures and parameters for reproducibility

5. Limitations and Considerations

When using non-AirGradient monitors, it’s essential to understand that sensor accuracy varies significantly between different technologies and manufacturers, which can impact the comparability and reliability of your data. These variations stem from fundamental differences in sensor design, manufacturing tolerances, and measurement principles, all of which affect how different sensors respond to the same environmental conditions.

These accuracy variations become particularly problematic when mixing different sensor types within a single monitoring network. Different sensors may exhibit systematic biases relative to each other, creating artificial spatial patterns in your data that reflect sensor differences rather than actual air quality variations. Calibration differences can compound this problem, and temporal drift rates may vary between sensor types, leading to data comparability issues that worsen over time.

Several strategies can help mitigate these data comparability issues. Cross-calibration involves calibrating all sensors against the same reference station, which helps reduce systematic differences between sensor types. Implementing comprehensive quality flagging systems that clearly mark data sources and quality levels helps users understand the limitations of specific measurements. Statistical adjustment techniques can be applied during post-processing to reduce systematic differences between sensor types, while maintaining transparency by documenting sensor types and known limitations ensures that data users understand the context and reliability of the information they’re accessing.

Long-term sustainability considerations become particularly important when selecting alternative monitoring platforms, as the long-term viability of your data and analysis capabilities depends heavily on the platform’s continued availability and support. Open source platforms generally offer better sustainability prospects due to community support that doesn’t depend on any single organization’s continued funding or business model. Commercial platforms, while often offering superior initial features and support, carry the risk of service discontinuation or pricing changes that could threaten project continuity.

Critical Success Factors:

• Sensor compatibility: Ensure chosen sensors have established calibration methods that lead to maximum reproducibility
• Data standardization: Implement consistent data formats and quality controls
• Platform stability: Choose platforms with strong long-term viability
• Community support: Leverage existing user communities for troubleshooting and best practices
• Documentation: Maintain detailed records of all procedures and decisions

6. Resource Directory

Successfully implementing air quality monitoring with alternative platforms requires access to comprehensive documentation, community support, and technical resources that can guide you through the specific challenges of your chosen system.

Platform-Specific Resources

Sensor.Community

• Main Website: https://sensor.community/en/
• Documentation: https://sensor.community/en/sensors/
• API Documentation: https://api-sensor-community.bessarabov.com/
• Community Forum: https://forum.sensor.community/
• GitHub Repository: https://github.com/opendata-stuttgart/sensor.community
• Data Archive: https://archive.sensor.community/

OpenAQ

• Main Website: https://openaq.org/
• API Documentation: https://docs.openaq.org/
• Data Explorer: https://explore.openaq.org/
• API Reference: https://api.openaq.org/
• GitHub Repository: https://github.com/openaq/openaq-api-v2

InfluxDB + Grafana

• InfluxDB Documentation: https://docs.influxdata.com/
• InfluxDB Air Quality Template: https://www.influxdata.com/influxdb-templates/air-quality-monitor/
• Grafana Documentation: https://grafana.com/docs/
• Grafana Air Quality Dashboard: https://grafana.com/grafana/dashboards/15524-air-quality/
• Getting Started Guide: https://grafana.com/docs/grafana/latest/getting-started/get-started-grafana-influxdb/

6.1. Data Analysis Tools

R Packages

• openair Package: https://openair-project.github.io/openair/
• The openair Book: https://openair-project.github.io/book/
• Scientific Paper: https://www.sciencedirect.com/science/article/abs/pii/S1364815211002064

Python Libraries

• pandas: https://pandas.pydata.org/
• matplotlib: https://matplotlib.org/
• seaborn: https://seaborn.pydata.org/
• numpy: https://numpy.org/
• plotly: https://plotly.com/python/
• scikit-learn: https://scikit-learn.org/
• Air Quality Analysis Tutorial: https://binh-bk.github.io/air-quality-analysis/

6.2. Community Forums and Support

General Air Quality Communities

• EPA Air Sensor Toolbox: https://www.epa.gov/air-sensor-toolbox
• Low-Cost Air Quality Monitors in Citizen Science: https://www.intechopen.com/chapters/1201982
• Stack Overflow: https://stackoverflow.com/ (tag: air-quality)
• Air Quality Management Exchange Platform: https://aqmx.org/content/air-quality-monitoring

Platform-Specific Communities

• Sensor.Community Forum: https://forum.sensor.community/
• OpenAQ Slack Community: Available through OpenAQ website registration
• Grafana Community: https://community.grafana.com/

Academic and Research Communities

• ResearchGate Air Quality Groups: https://www.researchgate.net/ (search: air quality monitoring)
• GitHub Air Quality Topics: https://github.com/topics/air-quality-data

6.3. Technical Documentation and Tutorials

Project Examples and Case Studies

• Air Quality with InfluxDB and Grafana: https://medium.com/@rubfi/monitoring-air-quality-open-data-63c9b12699b2
• Raspberry Pi Air Quality Monitor: https://blog.balena.io/build-an-environment-and-air-quality-monitor-with-raspberry-pi/
• Green City Solutions Case Study: https://grafana.com/blog/2022/10/04/visualizing-an-invisible-problem-how-grafana-helps-green-city-solutions-combat-air-pollution/

Educational Resources

• Python Air Quality Analysis: https://peyrone.medium.com/python-essentials-for-air-pollution-data-analysis-415cbbfca4d5
• Real-Time Air Quality Mapping: https://medium.com/@mahyar.aboutalebi/real-time-air-quality-mapping-in-4-easy-steps-python-3e4b7a09e2d3
• Environmental Data Analysis: https://peerdh.com/blogs/programming-insights/engaging-python-projects-for-analyzing-environmental-data

6.4. Government and Official Resources

International Organizations

• WHO Air Quality Guidelines: https://www.who.int/health-topics/air-pollution
• US EPA Air Quality Index: https://www.airnow.gov/aqi/aqi-basics/
• US EPA Sensor testing guidelines: https://www.epa.gov/air-sensor-toolbox/air-sensor-performance-targets-and-testing-protocols
• European Environment Agency: https://www.eea.europa.eu/themes/air

APIs and Data Sources

• AirNow API: https://docs.airnowapi.org/
• World Air Quality Index: https://waqi.info/

The key to success with alternative monitoring platforms lies in understanding that while the specific tools and procedures may differ from the AirGradient-focused approach described throughout this toolkit, the fundamental principles of effective air quality monitoring remain constant. Careful planning, meaningful community engagement, rigorous attention to data quality, and systematic impact measurement will ensure project success regardless of the specific hardware and software platforms you choose to implement.

By leveraging these resources and adapting the proven methodologies outlined in this toolkit, you can create an effective air quality monitoring project that serves your community’s needs while contributing to the global understanding of air quality challenges and solutions.

Developed by AirGradient in Collaboration with the UNDP Global Centre Singapore CC-BY-SA