Addressing Delhi Pollution: Predicting Air Pollution Level using Machine Learning
Increasing the level of Delhi Pollution is perilous, predicting air pollution levels using machine learning techniques and the role of IoT in blue sky analytics.
Delhi pollution was mentioned in the news,
Have you overheard that pollution level in Delhi spreading very quickly, yes, soo contaminated, how are people breathing their,
A curious me looked at Delhi, we are breathing here, really a polluted air, contaminated environment, an inquisitive me echoed in the news if and how I can contribute to controlling Delhi Pollution, hey listen! can’t we use the tardiest technologies in it, can we? Yes, we can!!!
In recent times, accelerated growth to industrialization and urbanization leads to improved lifestyle but it indirectly intensified air pollution in urban-regions considerably including Delhi. The Regularization and controlling of environmental pollution have withdrawn public scrutiny.
To the hark depth of Delhi and the world, let’s understand the numerous contamination added to intrinsic Delhi, the main cause of the increasing level of pollution in Delhi and how the freshest technology can constrain it with data analytics and prophecies.
Machine learning is the field where algorithms are designed to make a machine learn from the environment then readjust and behave accordingly. It learns from past experiences over time, improves and modifies accordingly to change and execute according to the environment. Hence, it would be the best-suited technique for predicting air pollution.
To make use of meteorological data can be featured for air quality predictions. It is easily available for urban regions and can yield desirable results. Parameters that incorporate in meteorological data are wind speed and direction, vertical wind speed, temperature, and relative humidity.
Delhi Pollution: Introduction
Delhi is one of the biggest and sprightly growing cities on the planet, home to more than 19.3 million people (according to recent projections), meeting the most terrible challenge that needs to address for this enormous and distinct population in terms of one of the basic life amenities: #cleanair. Recently, Delhi was granted the title, “World’s Most Polluted City” by WHO.
Those who are residing in polluted areas suffered from major health consequences due to air pollution, not even in Delhi but also in other places as well. Raised levels of dangerous and particulates gases are increasing the major risks of respiratory diseases such as heart disease, asthma, cancer, skin disease and more.
According to a survey conducted, every year most of the people die prematurely due to infected air, and so in Delhi. An average life span of life reduces continuously due to the increasing rate of air pollution.
These studies outline the undeniable need for air quality evaluation and control, i.e. Air Quality Index (AQI), it is elementally a number that informs how the air is polluted or contaminated at a given time for the specific area.
AQI forecast on time would be high in demand as it discloses the current trends in air quality and facilitates the local authorities to bring the correct measurements more productively and methodically.
Let’s have a look at critical parameters that participate in air pollution;
- Particulate matter(PM): Dust, Aerosols, and Soot
- Nitrogen monoxide (NO) and Nitrogen Dioxide (NO2 )
- Carbon monoxide (CO)
- Sulfur Dioxide (SO2), and Ozone (O3 )
Other pollutants also involved such as greenhouse gases, Carbon Dioxide (CO2), which also have an impact on both health and changes to the climate.
Delhi: Site Statistics description
New Delhi at 28.61ºN 77.23ºE is situated on the plains of river Yamuna having an elevation from 650ft to 820ft within the city.[1] From the past few years, it has become a hub of toxic air which is apparently more polluted air from the sea.
Also, the immense growth of urbanization, co-joining areas, development of residential, manufacturing and commercial industry have also placed that influentially making the air toxic in Delhi. It has been difficult to flush out the contaminated air due to this huge development, thereby increasing inner-city pollution.
The first prototype air purifier, Smog Tower (65 feet), was installed in Delhi to deal with air pollution recently. It is designed as such to serve air purifiers on a large scale. It comprises multiple filters in it which clean the air pollutants as air passes from it.
It is capable of filtering 6,00,000 cubic meters of air per day and can accumulate PM between 2.5 to 10. Carbon nanofibers are fitted at its peripheries as major components that focus on reducing PM load in the air.
Tech can revive in the blue sky analytics
As a Delhi resider, I was concerned by the infamous air pollution hovering over the city. Due to headways in technology, models or algorithms based on Artificial Intelligence are being extensively adopted for the purposes of predictions, many of them including air quality forecasting.
A machine learning passageway has considerably taken into account, diverse parameters are significantly verified and made sure for the revised forecasts rather than old and pure statistical models. In the conventional pollution operating method, data is assembled, examined and systematic and scientific instrumentation is carried out on them.
Among numerous machine learning intelligence-based proposals, Artificial Neural Networks(ANN) is the praiseworthy technique and has appeared the most deployed method for the predictions of air quality. The technique after estimated with Mean Square Error, Mean Absolute Error and R2 which proclaims Artificial Neural Network and Support Vector Regression are the best matches for predicting air quality of an area.
According to most of the studies conducted, many neural network-based hybrid models for air quality predictions, along with ANN algorithms PCA (Principle Component Analysis) have been practised to design models designed for the same concern, like, AdaptiveNeuro Fuzzy Interface System(ANFIS), PCAANN model, PCA_SYM model, and many more.
For example, by utilizing the Wavelet technique and Back Propagation Neural Network(W-BPNN) where the backpropagation neural network is revised using the wavelet-transform technique to foretell the concentration of SO2, NO2, and PM 10.
Along with it, different classification and regression techniques like Linear regression, SGD Regression, Random Forest Regression, Decision Tree Regression, Support Vector Regression, etc are the adaptive approaches to forecast the Air Quality Index of major pollutants such as SO2, NO2, PM 10, PM 2.5 and O3.
Game Changers: Turning of the Table
“You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.” — Buckminster Fuller
More advanced technology, easier to cross the barriers that might be an obstacle in improving the traditional methodologies.
Internet of Things is the game-changer that takes the lead, under its umbrella, it embraces the sensing devices, ML technology, and communication devices that together operate real-time air quality and implement analytics to develop better consciousness among people by alerting them on time.
Connected-devices have the features to measure exactly where the pollution is coming from, help in identifying causing factors that indirectly aid to IoT ecosystem to reduce air pollution.
Sensing devices are placed within the city to operate air quality and recovery data about contaminants at real-time conditions and allowing local government groups and authorities to take appropriate action and employ corrective measures before the AQI goes worse.
Additionally, factors that contributing to pollution are consistently monitored with IoT and AI, together they make it possible to combine inferences with past data to pinpoint patterns and decrease pollutants eventually. Hence, at a certain point, we need the Internet of Things(IoT) to monitor air pollution which is being contaminated at each and every second.
Along with the trendiest technologies, the government needs to consider laws and policies that benefit the climate and technological ecosystems, we know that AI brings with assorted risks such as efficiency and performance, privacy and security, economic and social factors, etc.
So, the government must ensure the safe and transparent cycle of technology, also an alliance of government with industrialists and policymakers must exhibit to relieve the challenges of AI and IoT and hold maximum benefits.
Conclusion
Air pollution in urban areas touches requisite options to restrict its deleterious consequences, several Machine Learning algorithms such as Regression and classification are highly deployed for controlling the “poor to worse” air pollution in Delhi and its adjoined places. As the machine learning applications are broadly designed for specific problems, so using the excellent ML technique is imperative that considers ecological and environmental factors to air pollution predictions.
Although Ml, AI, and IoT can help to create a sustained environment around us and proffer us a green future, it is the part of a process to attain the goal. One as an individual should not rely much on technologies, an organization, as well as humans, should fight against the root cause of air pollution, or even, try to do a bit from their side to minimize air pollution.
Originally published at https://www.analyticssteps.com on January 2020, 2020
