How can AI help prevent pollution peaks?
All around the world, urbanization acceleration and megalopolis development have well-documented consequences on ozone pollution. Another source of pollution stems from nitrogen dioxide (NO2). It is a gaseous air pollutant produced as a result of road traffic and fossil fuel combustion. Not only is it a pollutant, but its presence in the air contributes to the formation and modification of other air pollutants, such as ozone, particulate matter, and to acid rain. It has been scientifically proven that long-term exposure to NO2 levels as such currently observed in Europe may negatively impact lung function. It increases the risk of acute bronchitis, cough and phlegm, in particular in children.
Yet nitrogen dioxide is not the only byproduct of vehicle emissions, carbon monoxide (CO) is a toxic air pollutant also produced along the same process It can also be produced by a defectuous fireplace or heating system. Breathing CO at high concentrations leads to reduced oxygen transport by hemoglobin. This has health effects that include impaired reaction timing, headaches, lightheadedness, nausea, vomiting, weakness, clouding of consciousness, coma, and, potentially death. Another result of sunlight reaction on air containing hydrocarbon: ozone. This pollution too has direct consequences on the respiratory systems of exposed persons.
Artificial intelligence techniques can help meteorologists identify ozone pollution peaks, whether particulate matter, nitrogen dioxide, carbon monoxide or ozone, quicker to prevent and limit health risks for the exposed people.
As they specialize in weather prediction, meteorologists could benefit from predictive models to better forecast incoming meteorological and atmospheric variations. However, they are not Data Scientists and don’t have the required skills in Machine Learning and code to build these predictive models.
The data they work with daily are Small Data, as there is only a limited number of days when pollution peaks occurred. Traditional Machine Learning tools don’t handle well Small Data. MyDataModels offers TADA, an efficient tool to help meteorologists automatically create predictive models out of their Small Data datasets.
No data science training is needed to use TADA. Meteorologists can use their own data directly as they don’t have to normalize or preprocess it. TADA provides domain experts with convincing results in less than a minute on a traditional laptop.
MyDataModels brings a self-service solution for those who have Small Data and no data scientists.
Ozone pollution has major health consequences that can be managed by health authorities and meteorologists thanks to a better forecast of pollution peaks. Meteorologists could then alert public authorities to anticipate themoccurrence and implement the most relevant prevention plans.
“Predictive model allows professional to predict more efficiently future pollution peaks”
By gathering existing atmospheric data and adopting TADA, meteorologists can predict more efficiently future pollution peaks so that public authorities can organize better to lower the health risk for exposed populations. TADA will not replace meteorologists’ knowledge and expertise, but it can assist them predict, analyze and anticipate the major meteorological mutations of the 21st century.