Exploration of urban sustainability in India through the lens of sustainable development goals

The United Nations' (UN) Sustainable Development Goals (SDG) are a recognised metric for measuring environmental, economic, and societal progress. However, national or multinational-level analyses are more prevalent than sub-national types. The performance of 14 SDGs for 56 Indian cities (grouped into 6 regions) with the available 77 indicators (2020–2021) have been analysed. Pearson’s correlation, hierarchical clustering, data envelopment analysis, Theil index, etc. were used to infer existing status, interactions, inequality, efficiency, and interrelationships. Finally, policy suggestions have been offered coupled with limitations to mitigate the drawbacks of the Indian city SDG framework. The findings reveal the asynchronous nature of the SDGs. 18% of Indian cities register a poor track record of converting environmental performance into socioeconomic prosperity, while 55% of cities are lagging in performance compared to their respective states. Significant inequality exists among cities in various regions towards achieving the SDGs. The environment is adversely affected in a race to be economically powerful. So, mainstreaming the environment into development planning is urgently warranted.


Introduction
India has a great number of large cities, which add to the country's massive population.The country has 39 cities with populations of more than one million people.Mumbai and Delhi, two of these cities, have more than 10 million population.While these megacities add millions to the country's population, there are also smaller but densely inhabited cities, such as 396 cities with populations of 100,000 to 1 million people and 2,483 cities with populations of 10,000 to 100,000.Rural areas account for 65% of the population, while urban areas account for 35% of the population.This means the current (2021) rate of urbanisation is 2.1% [1].In numerous respects, cities have been viewed as critical to advancing the sustainable development goal.Cities have a profound and far-reaching impact on the environment and society beyond their borders due to their population density, economic prominence, wealth, and associated global resource requirements.Conversely, cities are centres of information, technology, and innovation, making them critical participants in any transition to sustainability.
The Sustainable Development Goals (SDGs) are a blueprint for long-term planning towards social, economic, and environmental well-being that all 193 UN member countries accepted in 2015.Transformation is required to address the SDGs' interrelated concerns.To establish egalitarian, inclusive, and sustainable ecosystems where all life may thrive, we must re-evaluate how societies work, how economies move, and how we engage with our planet.
Urban sustainability refers to practices and guiding ideas in urban planning that help us construct and improve our cities without permanently depleting our resources.The three pillars of sustainability are social sustainability (social well-being and health), economic sustainability (resource usage efficiency and economic payback), and environmental sustainability (resource use with environmental consequences).A sustainable city is planned with concern for its social, economic, and environmental effects (often referred to as the triple bottom line) as well as for the need to provide a stable home for its current inhabitants without compromising the ability of future generations to benefit from the same lifestyle.Sustainable cities place equal importance on the protection of the environment and the residents' economic, social, and physical well-being.
Cities all around the world should be learning from one another as they integrate the SDGs into existing planning processes, bridge data and communication gaps produced by administrative silos, and take the lead in identifying and resolving local needs to accomplish long-term, large-scale change.Downscaling objectives and indicators to the city level to help planning and policy in a local context, however, remains a difficulty.Very recently, the Intergovernmental Panel on Climate Change (IPCC) report [2] has given cognizance to the role of compound extremes and multiple climate change drivers operating in tandem in maximising disaster impacts in 12 Indian cities.Among the warnings, the intensity and frequency of hot extremes, such as warm days, warm nights, and heat waves, and decreases in the intensity and frequency of cold extremes, such as cold days and cold nights, are of severe importance for Indian cities. Important of these 12 cities that might go underwater are Mumbai, Mangalore, Cochin, Vishakhapatnam, Chennai, etc.Hence, this study has been conceptualised to understand and explore the contemporary conditions of Indian city SDGs, their interrelationship within SDGs as well as with some other city performance frameworks, SDG efficiency, and drawbacks, coupled with policy suggestions to mitigate them in the future.

Literature review
There have been a handful of works regarding this in the last few years.In one work [3], a co-production between researchers and local authority officials has been synthesised in five diverse cities: Bengaluru, Cape Town, Gothenburg, Greater Manchester, and Kisumu, for urban SDGs.Another study [4] tested the urban sustainable development goal 11 using five cities from Europe, Africa, and Asia.One work [5] has explored the relationship between data and governance regarding SDG 11 for Cape Town.Another work [6] has used SDG 11 to measure progress towards an inclusive, safe, resilient, and sustainable city, using German and Indian cities as case studies.A review [7] has focused on the evolution of indicators for monitoring sustainable urban development using SDGs and concluded that the SDG indicators provide the possibility of a more balanced and integrated approach to urban sustainability monitoring.A research study [8] has proposed a step-by-step process for implementing and integrating the SDGs in cities.Another study [9] has analysed sustainability on a local level (Romanian metropolitan area) by measuring 16 SDGs.In recent years, various reports have been published on the SDG assessment of municipalities or cities in Europe [10], the USA [11], Spain [12], Italy [13], Brazil [14], Bolivia [15], etc.A work [16] has reviewed the prospect of localising the SDGs from the urban resilience strategies of the 100 Resilient Cities (100RC) network and Cape Town.Another study [17] examined the extent to which the UN indicators will help cities evaluate their efforts to deliver sustainability and health outcomes.A group of authors [18] have proposed a system thinking approach towards several ongoing smart city initiatives with SDGs for the transition to a sustainable smart city using keyword-based search in the Scopus database.Another work [19] has done a content analysis of the intersections between 'urban' and 'equality' references across SDGs that can ensure 'leave no urban citizens behind' .A work [20] reflects on the SDGs agenda and Key Performance Indicators (KPIs) for sustainable and smart cities as a possible measurement tool for these multiple values, using Moscow as a case study.A work [21] has prepared a bibliometric literature review from the Web of Science (WoS) of 1991-2020 concerning urban sustainability assessment in the world.Some authors [22] have also composed a bibliometric literature review from the Web of Science of 1990-2020 regarding cities and SDGs and concluded that SDG monitoring and assessment of cities should take advantage of both consumption-based (footprint) accounting and benchmarking against planetary boundaries and social thresholds to achieve greater relevance for designing sustainable cities and urban lifestyles.Another work [23] has reflected SDG 'localization' derived from an action research project in Bristol.A work [24] has composed a comprehensive bibliometric analysis of 35 city labels to examine their (co-)occurrences during 1990-2019 from Scopus towards sustainable urban development.Another work [25] has emphasised accounting for the international spillovers of cities' SDG actions.A group of authors [26] have developed an analytical framework covering key components for local-level mainstreaming of the SDGs and then applied this to Shimokawa and Kitakyushu cities.Some recent works [27,28] have shown the interrelationships between environmental resource usage and socioeconomic development with globalisation, which is a key characteristic feature of urban areas.Other works [29,30] have reviewed city plans as well as peer-reviewed and grey literature to examine climate change adaptation action for 53 Indian cities with > 1 million population.They have established that 67% of these adaptation actions are merely in the implementation stage, i.e., a long way from achieving sustainability.There has been another work on national-level analyses involving SDGs [31].
Following these, several significant research gaps are identified: (a) There are no studies that use the full spectrum of the SDG framework for urban sustainability; (b) except for reports from the Sustainable Development Solutions Network (SDSN), almost all studies are composed of only one or a few of the SDGs; (c) most of the studies on urban sustainability are focused on one topic (e.g., climate change); (d) most of the existing studies only use prevailing analyses instead of incorporating new tools (e.g., data envelopment analysis); (e) majority of the studies for lower-middle highly populated rapidly urbanising countries like India encompasses the features the authors have mentioned earlier, etc.These deficiencies triggered us to compose a comprehensive study that can ensemble solutions for all of these scopes, even with the present data-scarce state of urban SDG analysis.This work is aimed at understanding the following from the perspective of major cities in India: • Achievements and shortfalls in terms of SDGs, • Interrelationships among SDGs, • Efficiency in utilising environmental scores towards socio-economic achievements, • Relative SDG performance • Interrelationships with other indices of performance, • Spatial inequality, • Policy suggestions to mitigate the potential drawbacks in the city SDG framework of India

Methodology
The SDG scores of 56 Indian cities have been collected from NITI Aayog [32].Out of 56, 44 urban units have a population of more than one million.The remaining 12 cities have populations of fewer than a million people.Three SDGs (viz., SDGs 14,15,and 17) have not been included, as the overall scoring of these SDGs is not available in the dataset.The 77 indicators included here cover various topics related to urban sustainability, such as clean energy, climate action, economic growth, education, forests, governance, health, industry, infrastructure, nutrition, poverty reduction, inequality reduction, urban development, water and sanitation, women's empowerment, etc. Six indicators related to Indian cities have been collected.These are: carbon footprint (CF) [33], population (pop) [34], city competitiveness index (CCI) [35], ease of living index (EoLi) [36], cost of living index (CoLi), and pollution index (PI) [37].
To understand the interrelationships between various SDGs, Pearson's correlation has been employed.For the assembly of correlation (via correlogram) between various SDG scores for every city included in this study, using the 'pheatmap' (version 1.0.12)packages with R (4.1.5),OriginPro 2021 has been used to create the heatmap of individual and grouped SDGs.
Along with this, the hierarchical clustering analysis (HCA) has been composed of two groups of features, viz., the environmental SDGs and the socioeconomic SDGs of the cities.The within-cluster sum of squared (WSS) method has been used, to find cluster numbers via the silhouette method through Euclidean distance using a single linkage.The silhouette method determines how well each point fits into its cluster and measures the quality of the clustering.The length of a line segment connecting two locations in Euclidean space is called the Euclidean distance.The 'cluster' (version 2.1.2),'dendextend' (version 1.15.2), and 'factoextra' (version 1.0.7)packages with R (4.1.5)have been used.
The performance assessment approach of data envelopment analysis (DEA) is used to determine the relative efficiency of decision-making units (DMUs).For a given level of socioeconomic development (represented here by socioeconomic SDGs), efficient cities use the fewest environmental resources (represented here by environmental SDGs), whereas inefficient cities use the most environmental resources for the same level of socioeconomic development (represented here by socioeconomic SDGs).The efficiency of a city can be assessed by comparing two environmental SDG inputs and 12 socioeconomic SDG outputs.Based on the applicability of our purpose, input-oriented DEA has been employed with the slack-based model (SBM) [38] and the variable return to scale (VRS) assumption, which minimises their inputs while maintaining consistent outputs, i.e., the same outputs with less input.According to the concept of returns to scale (RTS), efficient DMUs are categorised into three distinct zones: (a) increasing returns to scale (IRS) zone, (b) constant returns to scale (CRS) zone, and (c) falling returns to scale (DRS) zone.For these analyses, the 'deaR' (version 1.2.3) package with R (4.1.5)has been used.
The efficient frontier, which splits cities into two divisions, was used to explore a city's relative situation.Improvement goals have also been calculated for less efficient cities that are guided by more efficient cities, which can help us better understand the overall scope for improvements among cities in India as well as abroad.An efficient city, in this context, is one in which inputs are kept to a minimum, yet constant levels of success are achieved (outputs).The efficiency coefficient of each city (DMU), which ranges from zero to 1, is computed.DMUs with a one-to-one efficiency ratio are deemed efficient and form the efficiency frontier.The remaining DMUs (with an efficiency < 1) are considered inefficient, and targets for improvement can be assigned.The number of DMUs should be three times the sum of the inputs and outputs to have sufficient discriminating power [39].Another stipulation is that the number of DMUs equals the sum of the input and output variables.The input variables are two and the output variables are twelve in this study, and 56 DMUs meet both criteria, culminating in a model with sufficient discriminating power.
According to the returns to scale (RTS) concept, DEA can also categorise efficient DMUs into three distinct zones.DMUs can expand their outputs (here, socioeconomic SDGs) at a faster rate than their inputs in the rising returns to scale (IRS) zone (here, environmental SDGs).The input/output ratio (here, the environmental/socioeconomic SDG ratio) of DMUs is constantly maintained in the constant returns to scale (CRS) zone.In the decreasing returns to scale (DRS) zone, DMUs' inputs (here, environmental SDGs) are reduced more, while their outputs shrink considerably less (here, socioeconomic SDGs).This study is crucial because it reveals how difficult it is for DMUs (in terms of environmental SDGs) to improve their socioeconomic SDGs.If lambda sum = 1, DMU is in the CRS subzone; if lambda sum > 1, DMU is in the DRS subzone; if lambda sum < 1, DMU is in the IRS subzone [40].
Theil index [41] was used to quantitatively show trend disparities for each individual and group SDGs for the 56 Indian cities in terms of spatial variance.The Theil index is a crucial indicator of spatial disparity because it captures inequalities that information theory may decompose additively [42].It is calculated as: where N denotes the number of cities, X denotes the average score of SDGs, and X i denotes the score of SDG.
Due to the nature of the available city SDG dataset (single point, single year), most of the advanced analysis methods, like different types of regressions, future projections, etc., could not apply to this study.A flowchart of all the methodologies has been shown in Supplementary file 1.All the pertinent results related to this study are provided in Supplementary file 2.
For regional aggregation, six regions have been chosen.The distribution of 56 cities studied is as follows: central (23%), eastern (11%), north-eastern (12%), northern (18%), western (18%), and southern (18%) regions (see Supplementary file 1 Table S4, Fig. S1).When delving into this regional aggregation (Fig. 1c), it is seen that for environmental SDGs, southern and western cities perform better than the national average, while eastern, north-eastern, central, and northern cities' performance is lesser than that.For economic SDGs, western and central cities perform better than the national average.But northern, southern, north-eastern, and eastern cities perform less.For societal SDGs, southern, western, and northern cities perform better than the national average, while north-eastern, central, and eastern cities perform less.To sum up, Indian cities in the southern, western, and northern regions perform better than the national average, while cities in the central, north-eastern, and eastern regions perform less.Now, we understand the shortfalls of individual SDGs for Indian cities (Fig. 2a, b).Based on this, the lowest-performing cities with the highest cumulative gap are Dhanbad, Meerut, Itanagar, Guwahati, and Patna.The national average shortfall in Indian cities is 35.30.The cities with the highest lag in environmental SDGs would be Dhanbad, Amritsar, Agra, Ghaziabad, Faridabad, etc.The national average shortfall in Indian cities for environmental SDGs is 31.5.The cities with the highest lag in economic SDGs would be Dhanbad, Itanagar, Srinagar, Jodhpur, Kochi, etc.The national average shortfall in Indian cities for economic SDGs is 42.51.The cities with the highest lag in societal SDGs would be Meerut, Bhubaneshwar, Guwahati, Patna, Varanasi, etc.The national average shortfall in Indian cities for societal SDGs is 32.66.When delving into the SDG shortfall of city regions (Fig. 2c), it can be seen that the average national shortfall would be 35.31.Among 6 regions, 3 regions, viz., the southern (30.76), western (31.69), and northern regions (35.25), have a lesser shortfall than the national average.However, the remaining 3 regions, namely central (37.8), north-eastern (38.39), and eastern (40.04), have a higher shortfall.
The gap in the SDG score (i.e., inequality) of individual cities indicates if sustainable development has taken place synchronously or not.If different cities are assembled as per the increasing order of difference between the highest and lowest SDG scores, the SDGs would be: Composite SDG > 6 > 3 > 4 > 5 > 1 > 2 > 9 > 12 > 16 > 11 > 7 > 8 > 13 > 10.It can be seen that the top 5 SDGs with less inequality in the average SDG score for cities are combined SDGs 6, 3, 4, and 5. On the other hand, the bottom 5 are SDG 10, 13, 8, 7, and 11.These city SDGs have a large gap in scores among different cities.If cities of different regions are assembled as per the increasing order of difference between the highest and lowest SDG scores, the SDGs would be: Composite SDG > 5 > 3 > 8 > 9 > 16 > 6 > 4 > 1 > 10 > 14 > 12 > 2 > 7 > 11.Combined SDGs 5, 3, 8, and 9 have the lowest gap in regional SDG scores with one another.On the other hand, SDGs 11, 7, 2, and 12 have the highest differences among regions.From this, as per the decreasing order of regions with the highest shortfall, while comparing with another region, it would be: Eastern cities (for SDG 1, 3, 7, 8, 9, and Composite) > North-eastern cities (for SDG 6, 10, 11, 12, 16) > Central cities (for SDGs 2, 13, 14) > Northern cities (for SDG 5).Cities of the western and southern regions don't show the highest lag in any SDGs in comparison with cities from other regions of India.The authors have summed up the main results of the SDG performance of Indian cities with their respective regions (see Supplementary file 1 Table S5).Hierarchical clustering analysis (HCA) has been performed using two groups of data.First, the authors have used environmental SDG scores as input and economic SDGs and societal SDGs as output (Fig. 4a, also Supplementary file 1 Fig.S4).Second, a similar framework has been used, but individual SDG scores as input as well as output (Fig. 4b, also Supplementary file 1 Fig.S5).In both cases, the authors have determined the optimum number of clusters to be 5. From the first method, it can be seen on the left side that only 2 cities form the highest environmental performance cluster (skyblue colour), and only a handful form the cluster of lowest environmental performance (blue colour).The middle three clusters (orange, grey, and red), which can be interpreted as having better, intermediate, and lower environmental performance, are composed of an almost similar number of cities.This means, from the perspective of environmental SDGs, Indian cities fall into various categories with a similar number of members.On the right side, three clusters (orange, sky blue, and red) are formed by only a handful of cities, viz.4, 6, and 10 in number, respectively.The remaining two clusters (grey and blue) are composed of an almost similar number of members.A similar situation is seen in the second method (Fig. 4b), i.e., when individual SDG scores are considered instead of average grouping scores.These results indicate that higher environmental SDG performance does not correspond to better performance in economic and societal SDGs.This is evident when the performance of each city independently is tracked.Likewise, better performance in economic and societal SDGs does not mean that the cities have achieved similar environmental SDGs.

Efficiency
Data envelopment analysis (DEA) is employed next to assess the connecting efficiency of 56 Indian cities.For this, individual SDG scores from 56 cities have been used.Environmental SDGs were used as input and societal and economic SDGs were used as output.This exploration is focused on finding out the efficiency of Indian cities in translating better environmental features into socio-economic prospects, or lack thereof.The efficient and inefficient cities are 10 (i.e., 17.85%) and 46 (82.14%), respectively (Fig. 5a).
Linear combinations of indicator values in a group of comparable cities are used to generate improvement targets.Improvement targets point to the modifications that must be created to improve inefficient DMU efficiency.Peer cities are thought to be following best practices; therefore, inefficient cities should aim to emulate their behaviour as much as possible.Only a handful (n = 8; i.e., 14.28%) of cities have acted as peers ≥ 3 times (Fig. 5b).They, along with the times of appearance as references, are Amritsar (9), Dehradun (6), Coimbatore (6), Tiruchirappalli (3), Srinagar (3), Ghaziabad (3), Dhanbad (3), and Delhi (3).The inefficient cities concerning their efficient frontiers are shown here (Fig. 5c).
The idea of a return to scale provides insight into the environmental efficiency of DMUs' socioeconomic development (i.e., cities).It determines if the ratio of inputs (environmental SDGs) to outputs (socioeconomic SDGs) for a DMU is more productive or less productive.From the result, it is evident that only Jaipur belongs to the DRS sub-zone.This means Jaipur shows a decreasing socioeconomic return in terms of SDGs for more environmental input.Only 3 (i.e., 5.35%) cities (viz., Indore, Lucknow, and Ranchi) belong to the IRS sub-zone.This means these three cities show higher socioeconomic returns for environmental input.All the remaining (i.e., 52 or 92.85%) cities belong to the CRS sub-zone.This means a majority of Indian cities, at this stage, show an equal amount of socioeconomic return from environmental input.

Relative SDG scoring
We need to understand whether the SDG performance of Indian cities is better or worse than other comparable scores, such as-their state, national (Indian), regional (East & South Asia), income group (lower-middle income), or global scores.To understand this, a comparative index has been created of the relative SDG performance of cities.The relative performance of the city in SDG = (city score/other's score).This ratio can also be multiplied by 100 to convert the performance scale into a percentage.It has been done for each of the 56 cities for each SDG as well as the composite SDG score (see Supplementary file 1: Fig. S6).It is a general assumption that cities are responsible for the betterment of states or the forward advancement of states since districts have much fewer facilities than cities in terms of employment generation, access to education, health services, etc.Hence, in this case, most of the cities are supposed to perform better than their respective states.However, the results show a completely different picture.If the relative performance is categorised into 3 groups, it is clear that only 19 relative performance scores (i.e., 2.26%) in any one of the SDGs are similar to the state performance (= 1).Of the remaining, 460 relative performance scores (i.e., 54.76%) are worse performers than the state (< 1), while the remaining (i.e., 361 or 42.97%) are better than the state performance (> 1).From this, it can be interpreted that, in the case of most of the cities, other regions in their state (especially villages or smaller cities) must have performed better to bring up the average state SDG performance than these highlighted cities.The top 10 cities that have outperformed their state in a specific SDG would be Patna (SDG 4, 13), Ranchi (2, 13), Shillong (9), Dhanbad (2), Agartala (5), Kohima (9), Delhi (5), and Bhopal (9).The worst 10 cities that have lesser performance than their state in a specific SDG would be Kolkata (SDG 8), Kohima (10), Mumbai (8), Dhanbad (8), Shillong (8), Patna (8), Amritsar (13), and Tiruchirappalli (8).However, the authors think that the inclusion of more representative cities would have been helpful to conclusively understand this outcome.

Interrelationships with other city performance indices
In this section, the authors try to infer if any relationships exist between the SDG scores of Indian cities and some other indicators.Based on the availability of data, a heatmap has been composed ( From the correlogram (Fig. 6b), there are a few things to be seen.First, some relationships with a lesser degree of positive association exist.They are between CF-SDG, EoLi-PI, CoLi-PI, EoLi-CF, EoLi-Pop, PI-CCI, etc.Second, some relationships with an intermediate degree of positive associations exist.They are between SDG-CoLi, SDG-CCI, SDG-CoLi, EoLi-CoLi, etc. Third, some relationships with a higher degree of positive association exist.They are SDG-EoLi, EoLi-CCI, PI-CF, PI-Pop, CF-Pop, CF-CCI, CF-CoLi, Pop-CCI, Pop-CoLi, CCI-CoLi, etc. Fourth, only three negative correlations exist, which are SDG-CF, SDG-Pop, and SDG-PI, to an increasing degree.From the perspective of the relationships regarding city SDGs, it is easily understandable that the level of pollution index (PI), carbon footprint (CF), and population (pop) are negatively correlated with city SDGs.The pollution index is a measure of the city's overall pollution.Air pollution is given more weight than water pollution or accessibility, the two main pollution issues.Other types of pollution are given a low priority [40].So, higher pollution levels would hinder the achievement of environmental SDGs and, in turn, other socio-economic achievements.The carbon footprint of consumption would work on a similar path.An increase in population would trigger a higher requirement of environmental resources as well as lesser performance in environmental SDGs, which is coupled with a lack of or higher competition in access to goods and services related to socio-economic development, i.e., in societal and economic SDGs.The Ease of Living Index assesses the well-being of Indian city dwellers based on four main pillars: quality of life (35% weightage), economic ability (15%), sustainability (20%), and the citizens' perception survey (30%).It means the cities that are easier to live in would be placed higher in SDG scores.The City Competitiveness Index (CCI) measures the competitiveness of Indian cities across a variety of metrics.It employs Michael Porter's Diamond Model framework [44], which defines competitiveness as the sum of factor conditions, demand conditions, the backdrop for firm strategy and rivalry, as well as connected and supporting sectors.It is made up of four pillars that are divided into 12 sub-pillars to map all of the city's important dimensions.These four pillars are: factor conditions, demand conditions, the context for strategy and rivalry, and related and supporting industries.It can be understood that the more competitive the cities are, the better their SDG performance.The cost-of-living index (CoLi) is measured relative to New York City.It is a measure of the relative cost of consumer products, such as groceries, restaurants, transportation, and utilities.This means that inhabiting cities with overall better performance in the SDGs usually costs more.

Spatial variance of SDGs
The higher the Thiel index, the greater the difference for SDGs in each city.It is clear from the results of the individual SDGs (Fig. 7a) that the order of level of inequality is: 8 > 10 > 13 > 9 > 2 > 7 > 11 > 12 > 1 > 4 > 16 > 3 > 5 > 6.This shows that inequality is prevalent in the environmental and economic SDGs.As per the Theil index of grouped SDGs (Fig. 7b), it is evident that the level of inequality is highest in environmental SDGs, followed by economic SDGs, and then in societal SDGs.For policymakers, the priority of reducing inequality among major cities should be to focus first on the environment.

Discussion
This study has examined the performance, interrelationship, and efficiency of SDGs in 56 major Indian cities.To sum up the important results of this study, as for the SDG performance regarding environmental SDGs, as per the order of performance, southern > western > eastern > north-eastern > central > northern region cities.For economic SDGs, this order would be: western > central > northern > southern > north-eastern > eastern region cities.For societal SDGs, this order would be southern > western > northern > north-eastern > central > eastern region cities.For the overall SDG score, this order would be: southern > western > northern > central > north-eastern > eastern region cities.A significant degree of inequality exists among cities in various regions towards achieving the SDGs.Based on the inequality of various SDGs, the order would be eastern > north-eastern > central > northern region cities.From the Pearson correlation, it is clear that SDG 12 does not form any cluster with any other SDGs and is negatively correlated to all SDGs (excluding SDGs 10 and 11).Based on the number of positive correlations with any other SDGs, the order is 3 > composite > 7 > 5 > 4 > 1 > 16 > 10 > 6 > 13 > 11 > 8 > 2 > 9 > 12.This gives proof of the variable degree of synergy among various SDGs related to Indian cities. Due to the nature of clustering, accomplishing socio-environmental goals is hampered by economic goals in Indian cities, and vice versa.Environmental and social SDGs comprise a cluster that is separated from economic SDGs for both individuals and grouped SDGs.Higher environmental SDG performance does not equate to better success in economic and societal SDGs, according to the results of hierarchical clustering.Similarly, improved success in the economic and sociological SDGs does not imply that cities have achieved equivalent outcomes in the environmental SDGs.The results from the DEA clearly show that there are 17.85% of cities with a poor track record of converting environmental performance into socioeconomic prosperity.Furthermore, in their current state, the bulk of Indian cities exhibit an equal degree of socioeconomic return from environmental input (i.e., in the CRS zone).The authors think this proves a serious scope for decoupling economic growth from the environment, meaning the path towards better socioeconomic development must not come at an environmental cost.When the relative SDG performance scores of cities with their respective states in India are compared, it is evident that nearly 55% of cities are worse performers than the state.This suggests that other regions in Indian states (especially villages or smaller towns) must have performed better than these highlighted cities to improve the average state SDG performance.From the interrelationship with other performance scores, a few results come up about Indian cities.Higher levels of pollution and consumptive carbon footprints would obstruct the attainment of environmental SDGs as well as other socioeconomic goals.An increase in population would result in a greater demand for environmental resources and lower performance in environmental SDGs, which would be accompanied by a lack of or increased competition in access to goods and services related to socio-economic development, i.e., in societal and economic SDGs.SDG scores are higher in cities that are easier to live in.The better the SDG performance, the more competitive the cities are.It is frequently more expensive to live in cities with greater overall SDG performance.
While not exhaustive, the results came out intertwined with the thoughts provided in this work and could provide evidence-based observations on issues that should be considered for a more open and comprehensive process in the formulation and implementation of these global UN SDG objectives at the local level.The agendas' complexity and breadth, as well as their inclusive and participatory goals, necessitate an integrated governance approach that facilitates the formation of partnerships and dialogues between different levels of government (both horizontally and vertically within a single urban agglomeration), across sectors, and with various societal groups.Innovation and cross-sectoral collaboration are essential to achieving the aims of the SDG agendas.Only 8 years remain (2022-2030) to achieve the UN's SDG 2030 agenda (see Supplementary file 1: Fig. S8).The urban areas of India need special focus as the nature of the socioeconomic stage of development as well as ongoing and potential biophysical resource scarcity problems would affect the city dwellers harder.The authors think the singular focus on economic development is about bringing inequalities in access to various societal services, coupled with the scarcity of a range of biophysical resources, which are vital for the everyday lives of urban citizens.The disparity in SDG achievement for Indian cities is perceivable (see Supplementary file 1: Fig. S9).Local government and administrators should be able to understand critical aspects of a city's social, economic, and environmental performance through city-level sustainability assessments so that cities can be planned and managed to meet the needs of all residents while ensuring that environmental pressures do not exceed key thresholds.The SDGs can help with such judgements by offering a widely legitimate, goal-oriented framework and dashboard of objectives and indicators that encompass social inequality issues more comprehensively than prior sustainability assessment frameworks.

Limitations and policy recommendations
This work has a few limitations, which could also be viewed as prospects for future research as well as scopes for policy implementations coupled within.
First, there is a requirement for pertinent and reliable city SDG data management for Indian cities.The dataset prepared by NITI Aayog is full of data gaps, especially regarding individual indicators of SDG.Also, 3 SDGs (14, 15, and 17) are practically omitted from the dataset.SDG 14 (life below water) is completely absent.NITI Aayog has suggested that SDG 14 be excluded since it's only important for coastal areas.The authors suggest that when consumption-based impacts are measured (via footprint, LCA, etc.), major cities do have indirect yet significant connections with life below water (i.e., SDG 14).SDG 15 has only 2 indicators included, which don't even have rounded overall scores.NITI Aayog has also suggested that SDG 17 (partnerships for the goals) is not relevant at the urban local body level.The authors can also argue that to understand synergy and trade-offs, negative and positive feedback among SDGs 1-16 and SDG 17 is essential.This has hindered us in many ways to explore deeper into urban sustainability assessment.The authority with whom the data management has been entrusted (here, NITI Aayog) should resolve this.
Second, there is a necessity for a more comprehensive dataset.If the city SDG dataset of India is compared with that of Europe or the US city SDG dataset, a potential drawback could be found.For 4 SDGs, namely SDG 7, 9, 11, and 15, the number of indicators used for Indian city SDG indexing falls short by 1 (compared to the USA), 1 (Europe), 6 (Europe), and 2 (Europe) indicators (see Supplementary file 1: Fig. S7).It is a general understanding that robustness can be achieved through data abundance.Hence, it is suggested to incorporate more indicators, especially for those SDGs.
Third, there is the necessity of a more robust dataset (e.g., time series).The Indian city SDG dataset is composed of only one set of data.This specifically stops the temporal assessment as well as the future projection, which are of absolute necessity if India wants to comply with the UN SDG by 2030 for urban locations.
Fourth, this dataset only includes 56 cities among the nearly 400 cities inhabited by more than a million citizens.For many states, there is only one representative city, while there are dozens more.Though these cities might not be of national scale importance, they should be included to enrich the heterogeneity and representation of Indian cities in the SDG.
Fifth, the issue of the suitability of SDG for urban areas.At least 35% of urban citizens in India are living in slum areas (2018 data, [1]).Hence, to practically implement the 'leave no one behind' agenda in urban areas of India, the SDG framework should be an inclusive, equity-based measurement of SDG progress.The authors suggest that initiating people and authorities in these regions is of utmost importance if the Indian urban SDG doesn't want to be selectively applicable.
Sixth, the necessity is for co-creation, i.e., stakeholder dialogue and engagement.The goal of these urban SDGs is to prioritise performance over in-depth, locally relevant examinations of the causes of complex challenges like urban inequality and poverty.Each urban local government should choose an appropriate indicator set that is both realistic and feasible on the one hand and challenging and helpful in promoting its urban sustainability transition or even more substantive transformation on the other, ideally in consultation with its respective regional and national departments and ministries, as well as national associations of local governments.In a variety of circumstances, the UN-recommended SDG indicators may prove difficult to implement and ill-suited for local applicability.And here comes the scope of cocreating urban SDGs.
Seventh, there are philosophical challenges to urban SDGs.There is a clear risk that sectoral interests will take precedence over the agenda's longer-term objectives.Most local governments are still organised by sector, which makes it difficult to do the integrated, cross-cutting, and collaborative work required to achieve UN SDG agendas.
Eighth, the requirement of embracing complexity in the system framework.The potential conflicts, synergies, and trade-offs between the actions aimed at achieving the SDGs must be accounted for in these studies.This should include a discussion of the agenda's 'blind spots' , or subjects that aren't covered or aren't given enough consideration.
Ninth, a need for the adoption of transformational pathways.The ability to monitor and evaluate progress and alter the course of action as needed will play a role in realising the transformative potential of these agendas at the local level.On the other hand, when available, city governments can use their existing monitoring systems to supplement them with applicable and locally adapted SDG indicator frameworks.
Tenth, the necessity of integrating governance.A key aspect of the UN SDG is the integrated nature of sustainability, i.e., the importance of addressing the social, environmental, and economic dimensions of sustainability in unison.This requires multi-level collaboration and real-data-enriched and adaptive governance.It includes horizontal collaboration (between entities and actors at the same level), vertical collaboration (between actors at different levels, such as national, regional, and local), and collaboration among different types of actors.If needed, based on various features of cities, like culture, geographical location, employment generation, etc., each city can make its customised framework of laws and rules to implement and abide by.
Eleventh, is the issue of financing city sustainability.Local governments or higher authorities need to come up with feasible financial plans.These funds could emerge from the positive outcomes of various sustainability projects that are economically efficient and then be inserted again for betterment.City authorities can borrow funds from national or international funding agencies if need be.This aspect is of special significance for lower-middle-income economies, like India.

Conclusions
We think this work is the first of its kind, based exclusively on India, dealing with urban sustainability based on the UN SDG framework on a comprehensive scale.This study has many new contributions not seen in previous studies in similar urban sustainability research domains: (1) including almost all of the SDGs in the city sustainability framework instead of the traditional usage of 1-2 SDG indicators; (2) examining the links utilising different tools (i.e., analyses) than those commonly used, as evidenced by the literature; (3) exploring interrelationships among various SDGs at an individual city level; (4) interpreting the efficiency of Indian cities; (5) formulation of relative SDG performance measurement, not prevalent in the literature; (6) assessing inequality among individual and grouped SDGs; (7) exploring connections of some other performance indices with SDG scores, etc.
Cities are the primary drivers of the global consumption of goods and services; hence, the metrics utilised in SDG evaluation must be consumption-based.Cities' impacts or environmental threshold coherence offer significant wasted potential for influencing sustainable urban development.Interdisciplinary research is needed to measure, explain, and assist in alleviating the effects of urban consumption.This initial step necessitates ongoing collaboration among earth system, natural, environmental, system, and economic scientists to better understand the interlinkages among urban activities, consumption-based environmental footprints, and city planetary boundaries (for PB of Indian city Mumbai,

Fig. 1
Fig. 1 Achieving three types of SDGs for 56 Indian cities (a, b) and six regions (c).The SDG groups are environmental, economic, and societal SDGs.The average score of three SDG groups for six regions of Indian cities (c)

Fig. 2 Fig. 3
Fig. 2 Shortfalls in achieving SDGs for 56 Indian cities (a, b) and six regions (c).The average shortfall of individual SDGs for six regions (central, eastern, north-eastern, northern, southern, and western) of Indian cities (c) for the national average score

Fig. 5
Fig. 5 The efficiency of converting environmental SDGs into socioeconomic SDGs. a Grouping of efficient and non-efficient DMUs (cities, n = 56) of India; and Distribution of efficiency score for non-efficient DMUs (cities of India).b Ranking of efficient DMUs (cities) acting as peers (≥ 2 times) in reference sets c non-efficient DMUs (10 cities, red, inner circle) and their respective reference efficient DMUs (46 cities, green, outer circle).d Efficiency scores (0.74 to 0.93, here) for 10 inefficient cities in India Fig. 6a) to understand the distribution of data.It has also been coupled with clustering to infer associations among different indices.From the heatmap, it is seen that there are some data gaps, especially in CF, CoLi, CCI, and PI.The rest of the indices have a good abundance of data to infer relationships.Clustering has been applied based on the average linkage with Euclidean distance.The length of a line segment connecting two locations in Euclidean space is called the Euclidean distance.It can be seen that four indicators (viz., CCI, CoLi, PI, and EoLi) are closely associated with SDG in cities.Then, population (Pop) and CF are clustered in order.

Fig. 7
Fig. 7 Distribution of scores of inequality analysis based on Theil index.a inequality in individual SDG scores, and b inequality in grouped SDG scores among Indian cities economic SDGs is 32.5.The average performance of Indian cities in this economic SDG is 57.48.It means 46.42% of cities have lower economic SDG scores than the national average.The socially most performing cities are Kochi, Coimbatore, Tiruchirappalli, Panaji, and Shimla.The gap between the best (81.37,Kochi) and worst (53.25, Meerut) performing cities in the societal SDGs is 28.12.The average performance of Indian cities in these societal SDGs is 67.33.It means 46.42% of cities have lower societal SDG scores than the national average.The top 5 cities for environmental performance would be Kochi, Thiruvananthapuram, Shimla, Panaji, and Shillong.The gap between the best (95.5, Kochi) and worst (47, Dhanbad) performing cities in environmental SDGs is 48.5.The average performance of Indian cities on these environmental SDGs is 68.49.This means that 50% of cities have lower environmental SDG scores than the national average.Since societal and economic development are well connected and interlinked, they have been combined to yield socio-economic development.Likewise, when the same is performed, it yields the best socioeconomically performing cities, such as Shimla, Coimbatore, Pune, Chandigarh, Ahmedabad, etc.The gap between the best (72.68,Shimla) and worst (49.5, Dhanbad) performing cities in the societal SDGs is 28.12.The average performance of Indian cities in these socioeconomic SDGs is 62.4.It means 42.85% of cities have lower socioeconomic SDG scores than the national average.The best-performing cities are Shimla, Coimbatore, Chandigarh, Kochi, and Panaji in the composite score of all SDGs.The gap between the best (76, Shimla) and the worst (52, Dhanbad) performing cities in the composite SDG is 24.The average performance of Indian cities in this composite SDG is 64.65.This means 41.07% of cities have lower composite SDG scores than the national average.A comparative table has been summarised for different features of individual indicators of each SDG (see Supplementary file 1 TableS1).The descriptive statistics, based on individual SDGs and indicator data, have also been calculated (see Supplementary file 1 Tables | https://doi.org/10.1007/s43621-023-00158-2 1 3