Managing Air Quality Through Data Science (DASFA-MAQTDS)

 

 

Objectives and Scope

Databases and data science have become ubiquitous in several multi-disciplinary domains. Urban Informatics is a rapidly growing area for applications of data management techniques. In this regard, data-driven air quality assessment has assumed great importance as it can provide a strong basis for policy interventions aimed at improving public health. Although exposure to ambient air pollution is a major threat to human health in low-to-middle-income countries (LMIC) such as India and China, the challenges of data-driven air quality assessment become exacerbated in LMICs due to resource limitations and policy priorities focused on development and economic growth.

Countries such as India, China and Indonesia have some of the highest annual average ambient particulate matter exposure levels in the world. For example, the Global Burden of Disease study estimates that more than one million deaths every year in India are attributable to air pollution. Despite the poor air quality, the monitoring of air pollution levels is limited even in large urban areas in India and virtually absent in small towns and rural areas. Existing tools such as the SDG India Dashboard launched by NITI Aayog and MOSPI measure goals and targets at the national level. Although these provide data-driven insights to support development plans, negligible information is available at the town and city level. Studying and understanding air quality in urban spaces requires the collection and analysis of large volumes of heterogeneous, spatio-temporally varying data, including remotely sensed data from satellites, data from a range of ground-based sensors – weather stations, and networks of low-cost informal sensors—and historical and contemporary land use / mapping data.

 

It is worth noting that air quality (AQ) data satisfies all of the 5Vs of Big Data:

  1. Volume: AQ data is collected at relatively fine spatio-temporal granularities, and so is typically of huge scale;
  2. Velocity: The speed with which new AQ data gets generated is also very high – even a single low-cost sensor usually collects particulate matter concentration data at one-second intervals;
  3. Variety: AQ data comes from diverse sources such as dedicated static air quality monitors, mobile air quality monitors, satellites, existing imagery, thermal and emissions databases, among others, thereby implying a wide gamut of heterogeneity across data formats, including aspects of multi-modal data (e.g., text, image, video);
  4. Veracity: crowdsourcing of AQ data, coupled with issues concerning sensor reliability, and location & time uncertainty, veracity and reliability issues become critical; and
  5. Value: AQ data provides a key value proposition to municipal authorities in terms of ability to evaluate and assess policy interventions to improve air quality, and also adds value to the public at large (e.g., what is best time to jog?).


This will be the third workshop organized by Prof. Girish Agrawal centered on data challenges in assessing air quality. The first, “Data Challenges in Understanding the Urban,” was held as part of BDA 2020 on December 15, 2020. The second will be conducted as part of BDA 2021 in December 2021, and will focus on challenges in air quality data collection and curation. The DAFSAA workshop will be centered around efforts to develop an easy to use web and mobile information system which can serve as a visual tool to provide real-time, local information about outdoor air quality, and which can provide a simple assessment of the health impact of the same.

 


Target Audience

The workshop is envisioned as a space for academic researchers and practitioners to present and discuss their work in the core theme areas of Air Quality, and its dependence on urban morphology, as well as to introduce students to the key challenges and opportunities in studying the inter linkages between Air Quality, the UHI effect and Urban Morphology. This workshop will be useful for students and researchers of urban shape and form (morphology), architecture, urban planning, data analytics, mobile sensing, and low-cost sensor development, among other areas. In addition to academic researchers, the target audience includes urban planners, city administrators, and transport planners.