About this Guide

As we get a lot of questions about the science behind our air quality monitoring and how we manage to provide hourly air quality information at the street level, we’ve created this guide to explain the science behind our technology in the clearest terms possible.

By the end of the guide, you should understand why we developed a new approach to air quality reporting, how our technology works, and what makes us so different from other sources of air quality information.

A person can survive three weeks without food, three to five days without water, but only a few minutes without air.

 

Although the average person takes between 17,280 and 23,040 breaths a day, the health impacts of unhealthy air have been ignored for a long time. We think this is partly due to a lack of public awareness about air pollution, the invisible reality of the threat we’re dealing with, and limited access to real-time and actionable information to protect us.

 

The Shocking Statistics

 

According to the World Health Organization, more than 90% of the world’s urban population live in areas where air quality levels exceed safe recommended limits. Approximately 4.2 million deaths occur each year as a result of exposure to ambient air pollution, mainly from heart disease, stroke, COPD, lung cancer, and acute respiratory infections. In 2020, 9-year-old Ella Kissi-Debrah also became the first person in the world to have air pollution listed as a cause of death after suffering from a fatal asthma attack.

 

A Man with a Mission

 

Back in 2014, BreezoMeter’s CEO & Co-founder Ran Korber - an environmental engineer and husband of an asthma sufferer with a child on the way - recognized the scale of the problem and lack of available solutions. He asked the following question: What if technology enabled us to forecast pollution just like we can forecast the weather?

 

And so, BreezoMeter was born.

 



 

Air quality reporting has been around for a long time (since the 1940s). However, when BreezoMeter’s early scientists and engineers scrutinized readily available sources of information, we found this information wasn’t suitable for individuals looking to manage their exposure to air pollution.

 

Here's why:

 

A) Government Monitoring Stations

 

The most common sources of freely available air quality information are monitoring stations deployed by governments- they look like walk-in rooms containing many different measurements. Monitoring stations are set up in specific locations and house differing numbers of monitors depending on the particular station. Each monitor measures just 1 pollutant and they are extremely expensive pieces of equipment (costing up to $150,000 per station).

 

The Trouble with Station Reporting

 

While the information provided by government monitoring stations is extremely reliable, many don’t report in real-time, and no monitoring station is able to provide forecasts. They also require many stations to be dispersed across wide areas for accuracy across the board.

 

For these reasons, government monitoring stations fail to deliver a truly personalized picture of air quality exposure in real-time. This is because air quality changes on an hourly basis and at the street level. It’s also important to remember that different individuals may be more or less at risk from specific pollutants depending on their demographic profile (age, health status, etc). This presents a further challenge if the closest monitoring station reports on only a limited number of pollutants.

Lastly, governments rely on physical sensors, which are frequently taken offline or damaged during wildfire events  - leaving individuals without reporting when they need it the most.

 

B) Satellites

 

Earth-observing satellites are great for providing an aerial view of air quality. However, most satellites scan by orbiting the earth. This means their reporting isn’t live, because they only cross-specific locations at specific times of the day. Visual obstructions like cloud cover can also serve to mask the view of pollution from above, impacting the reliability of some of the air quality data provided by satellites.

 

Satellite

C) Low-Cost Sensors

 

Low-cost personal sensors are becoming more and more popular, ranging in type from stationary to portable.

 

At best, these sensors provide an excellent means of air quality monitoring for specific pollutants. However, each sensor still only reports on one pollutant, and can’t provide forecast information. The measurements provided by low-cost/personal sensors are also commonly impacted by levels of Relative Humidity and temperature and require regular calibration because they naturally tend to drift over time.

 

Without the ability to conduct in-depth air quality data analysis or advanced QA, or to check and modify the calibration over time, the real-time accuracy of these sensors is questionable.

 

Guide to Evaluating an Air Quality Data Provider

 

Empowering Individuals & Businesses

 

BreezoMeter’s scientists resolved to develop a solution that would enable people to decrease their exposure to polluted air: By knowing what the air quality picture was going to look like at their location, individuals would be able to plan better alternative routes to or from work and home, know when and how to use indoor air purifiers, and even move regular jogging routes to a safer location.

By leveraging this new air quality information, several industries could also work towards creating healthier and engaging product experiences for their customers: Automotive manufacturers could create better air filters and protect cabins from pollution, HVAC manufacturers could develop systems that responded to the environmental reality, and healthcare providers could deliver timely warnings and advice to at-risk patient groups.

 

Creating the Algorithm


BreezoMeter’s researchers and engineers set out to develop a proprietary algorithm for air pollution prediction, capable of offering never before seen accuracy and certainty - but the team faced a number of challenges:

Challenge 1: Ensuring Reliability

 

The first challenge related to the provenance of the data incorporated in the analysis. Whereas other applications for air quality analysis use data from unverified sources, such as smartphones, individual reports, and off the ground report sources such as satellites, BreezoMeter wanted to verify all information in real-time and ensure the sources used were reliable.

 

Challenge 2: Computational Power


Atmospheric circulation requires huge computational power for accurate predictions, especially in models with high spatial resolution such as those incorporated in BreezoMeter’s technology. This presented another challenge.

 

The Solution: Weather-inspired Modeling


The final answer to the problem of forecasting the wave motion of pollution was inspired by global and regional weather models, used by meteorologists to track the progress of weather and make predictions. The expertise of our R&D team in the areas of atmospheric science, turbulent flow, and convection meant we could apply scientific principles and mathematical modeling to define how air pollution algorithms should behave.

To accurately represent the spatial representation of air pollution, BreezoMeter calculates the dynamic nature of air pollution dispersion between the monitoring stations:

 

 

1. Moving Beyond the Sensors


BreezoMeter factors information taken from multiple data sources in addition to the information provided by government monitoring stations. In total, this equates to information from more than 14,000 data sources around the world.

 

Information from monitoring stations, low-cost sensors, satellites, meteorological data, live traffic, land cover information & more are combined to increase the accuracy of prediction, together with air quality dispersion models (some for natural and some for human-caused sources of emission).

 

Group 4425

2. Processing & Validation


To ensure the raw data we collect is reliable, all the information goes through a process of organization and validation before it is inputted into our models. Each new registered value is tested by several methods to confirm it falls within the normal trend and range and is not a false data point. The data is also normalized and invalid outliers are eliminated, which can sometimes arise from sources such as uncalibrated or faulty monitoring stations.

 

3. Predicting Air Quality


Similar to climate and weather forecasting models, BreezoMeter is a big data platform that uses physical equations, statistical analysis, and data collected in the field to predict and track the evolution of air pollution through inhabited areas. Our air quality modeling consists of hundreds of calculations at once, which enable us to calculate air pollution on an hourly basis and at the street level.

 

4. QA & Ensuring Accuracy

 

We constantly check the accuracy of our reporting in a number of ways:

a) We continuously scan for unrealistic anomalies that are likely to be caused by malfunctions or offline sensors.

b) We compare our data retrospectively to monitoring station data when it becomes available.

c) We constantly record our level of prediction error and report on this.

 

Learn More about Our Accuracy Validation



In order to provide accurate air quality information during wildfire events, we’ve incorporated a sophisticated smoke model into our reporting.

The inputs for these models include:

a) Information from satellites that measure different light bandwidths to detect the presence of fire on earth, the stage of fire & the type of Particulate Matter that is emitted

b) Land cover information to understand the type of vegetation that is being burned.

c) Meteorological conditions like rain and wind to learn about the direction and amount of smoke in a particular area.

d) Chemical processes that might alter the types of pollutants, and pollutant levels downwind of a fire.

BreezoMeter's Smoke Model

We partner with several different traffic data providers who deliver real-time information about the number of car lanes at a destination, average speed, the severity of reported traffic jams, and more.


We collect this data every 12 minutes for each section and use unique machine learning algorithms which take into account the parameters and geographical location to calculate the pollution emission of each traffic jam to its local surroundings.

Our air quality grid is essential for understanding our 5-meter granularity. We like to think of this as a kind of woven blanket laid over the world: Each tiny square is made up of tiny grid points that are just 500 meters or less away from each other:

 

BreezoMeter's Air Quality Grid

 

When users query the air quality at their location, they receive information based on the closest grid point.

 

Before we added our traffic vectors model, we used to provide air quality information at the resolution of 500 meters around the world. In itself, this is impressive - but with the help of live traffic information, we can now calculate information within the squares of the grid itself:

1. User queries a location
2. We check if a traffic jam is registered at that location - alongside all other air quality factors based on the query’s closest grid point.
3. We report air quality in real-time at the resolution of 5 Meters/16.5 Feet.

 

Group 4457 (1)

 

BreezoMeter is currently the only provider that can deliver air pollution at this level of granularity!

 

As the air quality indexes used by different countries adopt completely different approaches to air quality reporting, we had to develop a way to ensure our air quality information could be understood by anyone, anywhere.

To solve this problem, we provide actionable and health-focused air quality information for approximately 60 different national AQIs around the world.

BreezoMeter’s Air Quality Index (BAQI)

 

BreezoMeter Air Quality Index


Our scientists have also created a standardized universal index for reporting on air quality which individuals can use to understand air quality wherever they are.

We based the BreezoMeter Air Quality Index (BAQI) on academic research and on different health-based AQIs around the world and ensured our definitions of ‘unhealthy air’ were based on actual impact to health.

The BAQI means that people can now refer to one source of truth for understanding and comparing air quality.

 

To deep-dive further into our BAQI index, read our dedicated guide. 

 

 

 

BreezoMeter’s technology and the methodology behind our continuous accuracy validation have been tested by unaffiliated and authoritative bodies - this is something no other air quality data provider can demonstrate.

1. Air Quality Consultants conducted a critical review of our data and produced a detailed scientific report supporting the accuracy behind the model.

2. Google tested our technology and claims "By using sophisticated algorithms to calculate air pollution, BreezoMeter has quickly established itself as the world leader in hyper-local air quality data".

3. The world's biggest brands have tested and chosen BreezoMeter to be their official source of air quality information - including Apple, Verizon, Boehringer Ingelheim, L’Oreal, and many more.

 

 

BreezoMeter's Partners

 

Group 4496

See Real Case Studies

Talk to Us about Integration