urban waste analytics - wiscwarwick · 2016. 5. 2. · urban waste analytics using data analytics...

1
Urban Waste Analytics Using data analytics and citizen science to improve waste understanding and operations Urban waste is one of the most pressing issues facing our growing cities globally. According to studies, global production of waste will reach 11 million tons per day by 2100 [1]. Currently at 2.2 million tons per day, the amount of waste produced is becoming unmanageable [2]. As landfills and incinerators multiply, disposal of municipal, medical, electronic and industrial wastes have become financial and political burdens for municipalities, while posing risks to the health and well-being of the planet and its inhabitants. Many cities invest in recycling to extract value from waste streams and reduce disposal costs while, simultaneously, cheap labor and relaxed environmental regulations has led to the development of recycling economies in the developing world [3]. Tackling the issue of consumption, waste production, global flows and disposal requires a fundamental and integrated understanding of the waste eco-system. Related Work Previous work in forecasting waste generation has been explored through a variety of modeling methods. Modeling methodologies have included group comparison, correlation analysis, multiple regression analysis, input–output analysis, time-series analysis and system dynamics modeling. These models focus on identifying the underlying relationship between variables which drive waste generation [4]. At the municipal level, Oribe-Garcia (2015) [5] identified urban morphology, tourism activity, level of education and income as the most influencing factors leading to municipal solid waste (MSW) generation, while Daskalopoulos (1998) [6] used single regression analysis to link gross domestic product and related total consumer expenditure as strong correlating factors in waste generation at the country level. Navarro-Esbri et al. (2002) [7] and Rimaityte (2012) [8] have used traditional time-series approaches such as Autoregressive and Integrated Moving Average and seasonal Autoregressive and Integrated Moving Average (sARIMA) to predict generation. Xu et al. (2013) [9] disregarded demographic and socioeconomic factors and forecasted waste generation using hybrid sARIMA model and grey system theory. Recent approaches have explored data-driven models which have the advantage of learning the relationship between variables from supplied data in order to predict future values. Zade et al. (2008) [10] used artificial neural networks to predict weekly waste generation in Tehran while Abbasi (2013) [11] used a combination of partial least square for feature selection and support vector machine modeling to predict for the same area. Aim of this Research This research aims to explore and untangle the complex relationships between society, cities and waste through quantifying,modeling,predicting,evaluating and making visible the various waste processes that exist in cities. In particular, this research focuses on New York City’s municipal waste stream as well as the city’s electronic discards known as e-waste. Research aims include: 1) leverage historical data to predict waste generation in order to improve the New York City’s Department of Sanitation operations; 2) model waste generation at small spatial scales to understand generation patterns; 3) explore public participation in waste research as an alternate form of outreach and education. Initial Prediction Results Historical municipal solid waste (MSW) collection data supplied by the New York City Department of Sanitation was used in conjunction with other datasets related to New York City, including population, weather and tax lot data. A Gradient Boosting Regression Model was built and trained on this historical data from 2005-2011. The model is able to forecast weekly MSW generation tonnages for each of the 232 geographic sections and across three waste streams with high accuracy and is able to adapt to fluctuations in generation caused by holidays, special events, seasonal variations and weather related events [12]. Prediction Visualisation Visualization is an important component to understanding and further exploring data. Figure 1 is an interactive visualization of waste generation, disaggregated to the NYC tax lot level using a population estimate as a proxy for generation. Users can navigate the map to examine individual lot generation and the historical collection data of that lot [12]. Research Impact Societies across the globe and throughout history have produced waste. However, it is only within the last half century that the type and quantity of waste has changed dramatically and continue to do so. This makes urban waste analytics a rich area of exploration not only because of the potential to understand society’s phenomenological patterns of waste generation, the interdisciplinary nature of waste or the global interconnectedness via waste, but also because data-driven insights can lead to practical and fundamental day-to-day waste management improvements. Furthermore, this research is applicable, extendable and replicable for the growing cities across the globe. We are grateful to the City of New York for their collaboration and support of this reserch. References [1] Hoornweg, Daniel, Perinaz Bhada-Tata, and Chris Kennedy. “Waste production must peak this century.” Nature 502.7473 (2013): 615-617 [2] Hoornweg, Daniel, and Perinaz Bhada-Tata. “What a waste: a global review of solid waste management.” (2012). [3] Alexander, Cathering, and Joshua Reno. Economics of recylcing. Catherine Alexander & Joshua Reno. Zed Books, 2012. [4] Beigl, Peter, Sandra Lebersorger, and Stefan Salhofer. “Modelling municipal solid waste generation: A review.” Waste management 28.1 (2008): 200-214. [5] Oribe-Garcia, Iraia et al. “Identification of influencing municipal characteristics regarding household waste generation and their forecasting ability in Biscay.” Waste Management 39 (2015): 26-34. [6] Daskalopoulos, et al,. “Municipal solid waste: a prediction methodology for the generation rate and composition in the European Union countries and the United States of America.” Resources, conservation and recycling 24.2 (1998): 155-166. [7] Navarro-Esbrı, J, et al,. “Time series analysis and forecasting techniques for municipal solid waste management.” Resources, Conservation and Recycling 35.3 (2002): 201-214. [8] Rimaitytė, Ingrida et al. “Application and evaluation of forecasting methods for municipal solid waste generation in an eastern- European city.” Waste Management & Research 30.1 (2012): 89-98. [9] Xu, Lilai et al. “A hybrid procedure for MSW generation forecasting at multiple time scales in Xiamen City, China.” Waste management 33.6 (2013): 1324-1331. [10] Jalili Ghazi Zade, M, and R Noori. “Prediction of municipal solid waste generation by use of artificial neural network: A case study of Mashhad.” (2007). [11] Abbasi, M et al. “Forecasting municipal solid waste generation by hybrid support vector machine and partial least square model.” International Journal of Environmental Research 7.1 (2012): 27-38. [12] http://www.opentrashlab.com/ [13] http://crowdcrafting.org/project/landfill/ Warwick Institute for the Science of Cities The University of Warwick, CV4 7AL wisc.warwick.ac.uk “The average New Yorker throws out nearly 15 pounds of waste a week, adding up to millions upon millions of tons a year. To be a truly sustainable city, we need to tackle this challenge head on.” Bill de Blasio Mayor of New York City There is so much data on the waste eco-system that visualisations are vital in understanding it Nicholas Johnson PhD 2014-2021 Image from Landfill Hunter, a crowdsourcing project to landfill size and locations across the United States

Upload: others

Post on 19-Feb-2021

2 views

Category:

Documents


0 download

TRANSCRIPT

  • Urban Waste AnalyticsUsing data analytics and citizen science to improve waste understanding and operations

    Urban waste is one of the most pressing issues facing our growing cities globally. According to studies, global production of waste will reach 11 million tons per day by 2100 [1]. Currently at 2.2 million tons per day, the amount of waste produced is becoming unmanageable [2]. As landfills and incinerators multiply, disposal of municipal, medical, electronic and industrial wastes have become financial and political burdens for municipalities, while posing risks to the health and well-being of the planet and its inhabitants. Many cities invest in recycling to extract value from waste streams and reduce disposal costs while, simultaneously, cheap labor and relaxed environmental regulations has led to the development of recycling economies in the developing world [3]. Tackling the issue of consumption, waste production, global flows and disposal requires a fundamental and integrated understanding of the waste eco-system.

    Related WorkPrevious work in forecasting waste generation has been explored through a variety of modeling methods. Modeling methodologies have included group comparison, correlation analysis, multiple regression analysis, input–output analysis, time-series analysis and system dynamics modeling. These models focus on identifying the underlying relationship between variables which drive waste generation [4].

    At the municipal level, Oribe-Garcia (2015) [5] identified urban morphology, tourism activity, level of education and income as the most influencing factors leading to municipal solid waste (MSW) generation, while Daskalopoulos (1998) [6] used single regression analysis to link gross domestic product and related total consumer expenditure as strong correlating factors in waste generation at the country level. Navarro-Esbri et al. (2002) [7] and Rimaityte (2012) [8] have used traditional time-series approaches such as Autoregressive and Integrated Moving Average and seasonal Autoregressive and Integrated Moving Average (sARIMA) to predict generation. Xu et al. (2013) [9] disregarded demographic and socioeconomic factors and forecasted waste generation using hybrid sARIMA model and grey system theory.

    Recent approaches have explored data-driven models which have the advantage of learning the relationship between variables from supplied data in order to predict

    future values. Zade et al. (2008) [10] used artificial neural networks to predict weekly waste generation in Tehran while Abbasi (2013) [11] used a combination of partial least square for feature selection and support vector machine modeling to predict for the same area.

    Aim of this Research

    This research aims to explore and untangle the complex relationships between society, cities and waste through quantifying, modeling, predicting, evaluating and making visible the various waste processes that exist in cities. In particular, this research focuses on New York City’s municipal waste stream as well as the city’s electronic discards known as e-waste.

    Research aims include:

    1) leverage historical data to predict waste generation in order to improve the New York City’s Department of Sanitation operations;

    2) model waste generation at small spatial scales to understand generation patterns;

    3) explore public participation in waste research as an alternate form of outreach and education.

    Initial Prediction ResultsHistorical municipal solid waste (MSW) collection data supplied by the New York City Department of Sanitation was used in conjunction with other datasets related to New York City, including population, weather and tax lot data. A Gradient Boosting Regression Model was built and trained on this historical data from 2005-2011. The model is able to forecast weekly MSW generation tonnages for each of the 232 geographic sections and across three waste streams with high accuracy and is able to adapt to fluctuations in generation caused by holidays, special events, seasonal variations and weather related events [12].

    Prediction Visualisation

    Visualization is an important component to understanding and further exploring data. Figure 1 is an interactive visualization of waste generation, disaggregated to the NYC tax lot level using a population estimate as a proxy for generation. Users can navigate the map to examine individual lot generation and the historical collection data of that lot [12].

    Research ImpactSocieties across the globe and throughout history have produced waste. However, it is only within the last half century that the type and quantity of waste has changed dramatically and continue to do so. This makes urban waste analytics a rich area of exploration not only because of the potential to understand society’s phenomenological patterns of waste generation, the interdisciplinary nature of waste or the global interconnectedness via waste, but also because data-driven insights can lead to practical and fundamental day-to-day waste management improvements. Furthermore, this research is applicable, extendable and replicable for the growing cities across the globe.

    We are grateful to the City of New York for their collaboration and support of this reserch.

    References[1] Hoornweg, Daniel, Perinaz Bhada-Tata, and Chris Kennedy. “Waste production must peak this century.” Nature 502.7473 (2013): 615-617[2] Hoornweg, Daniel, and Perinaz Bhada-Tata. “What a waste: a global review of solid waste management.” (2012).[3] Alexander, Cathering, and Joshua Reno. Economics of recylcing. Catherine Alexander & Joshua Reno. Zed Books, 2012.[4] Beigl, Peter, Sandra Lebersorger, and Stefan Salhofer. “Modelling municipal solid waste generation: A review.” Waste management 28.1 (2008): 200-214.[5] Oribe-Garcia, Iraia et al. “Identification of influencing municipal characteristics regarding household waste generation and their forecasting ability in Biscay.” Waste Management 39 (2015): 26-34. [6] Daskalopoulos, et al,. “Municipal solid waste: a prediction methodology for the generation rate and composition in the European Union countries and the United States of America.” Resources, conservation and recycling 24.2 (1998): 155-166.[7] Navarro-Esbrı, J, et al,. “Time series analysis and forecasting techniques for municipal solid waste management.” Resources, Conservation and Recycling 35.3 (2002): 201-214.[8] Rimaitytė, Ingrida et al. “Application and evaluation of forecasting methods for municipal solid waste generation in an eastern-European city.” Waste Management & Research 30.1 (2012): 89-98.[9] Xu, Lilai et al. “A hybrid procedure for MSW generation forecasting at multiple time scales in Xiamen City, China.” Waste management 33.6 (2013): 1324-1331.[10] Jalili Ghazi Zade, M, and R Noori. “Prediction of municipal solid waste generation by use of artificial neural network: A case study of Mashhad.” (2007).[11] Abbasi, M et al. “Forecasting municipal solid waste generation by hybrid support vector machine and partial least square model.” International Journal of Environmental Research 7.1 (2012): 27-38.[12] http://www.opentrashlab.com/[13] http://crowdcrafting.org/project/landfill/

    Warwick Institute for the Science of CitiesThe University of Warwick, CV4 7AL

    wisc.warwick.ac.uk

    “The average New Yorker throws out nearly 15 pounds of waste a week, adding up to millions upon millions of tons a year. To be a truly sustainable city, we need to tackle this challenge head on.”

    Bill de Blasio Mayor of New York City

    There is so much data on the waste eco-system that visualisations are vital in understanding it

    Nicholas JohnsonPhD 2014-2021

    Image from Landfill Hunter, a crowdsourcing project to landfill size and locations across the United States