ArticlePDF Available

Abstract and Figures

An overview of the "COINr" R package, which is for building and analysing composite indicators.
Content may be subject to copyright.
COINr: An R package for developing composite
William Becker 1, Giulio Caperna2, Maria Del Sorbo 3, Hedvig Norlén1,
Eleni Papadimitriou2, and Michaela Saisana 2
1Freelance consultant, Ispra, Italy 2European Commission, Joint Research Centre, Italy 3European
Innovation Council and SMEs Executive Agency, Belgium Corresponding author
DOI: 10.21105/joss.04567
Editor: Mehmet Hakan Satman
Submitted: 06 July 2022
Published: 11 October 2022
Authors of papers retain copyright
and release the work under a
Creative Commons Attribution 4.0
International License (CC BY 4.0).
Composite indicators (CIs) are aggregations of indicators that aim to measure complex, multi-
dimensional and typically socio-economic concepts such as sustainable development (Conceição,
2020), innovation (Dutta et al., 2020), globalisation (William Becker et al., 2021), gender
equality (Equal Measures 2030, 2019), and many more. CIs are very widely used in policy-
making, by international organisations, and they are equally well-covered in academic literature
(El Gibari et al., 2019;David Lindén et al., 2021;Stefana et al., 2021). They are often used to
rank and benchmark countries or regions to help direct policy making, but are also frequently
used for advocacy (Cobham et al., 2015).
The COINr package, introduced in this article, aims to provide a harmonised development
environment for composite indicators that includes all common operations from indicator
selection, data treatment and imputation up to aggregation, presentation of results and
sensitivity analysis. COINr enables development, visualisation and exploration of methodological
variations, and encourages transparency and reproducibility.
Statement of need
Existing tools
Some dedicated tools for composite indicators exist: in Microsoft Excel, the COIN Tool is a
spreadsheet-based system which allows users to build and analyse a composite indicator (W.
Becker et al., 2019). In MATLAB, there are some packages addressing specic parts of index
development: the CIAO package uses a nonlinear regression and optimisation approach to
tune weights to agree with expert opinions (D. Lindén et al., 2021). In R (R Core Team, 2022)
there is an existing package for composite indicator development, called compind (Fusco et
al., 2018), focusing on weighting and aggregation, although this gives no special consideration
to hierarchical structures, uncertainty and sensitivity analysis, and so on.
The Python library CIF gives a number of tools for building composite indicators, from loading
data to aggregation and visualisation (Vraná, 2022). This is focused in particular on Business
Cycle Analysis. Finally, there is a recently launched Web-based tool called the MCDA Index
Tool (Cinelli et al., 2021). This is mostly focused on multi-criteria decision analysis, and
doesn’t include dierent levels of aggregation.
COINr is a signicant step beyond existing composite indicator tools in many respects. COINr
wraps all composite indicator data, analysis and methodological choices into a single S3 class
Becker et al. (2022). COINr: An R package for developing composite indicators. Journal of Open Source Software,7(78), 4567. https:
object called a “coin”. A coin is a structured list including:
Indicator data sets for each processing step (e.g. imputation and normalisation).
Metadata pertaining to indicators and units (e.g. names and weights, but also the
hierarchical structure of the index).
A record of the COINr functions applied in constructing the coin.
This enables a neat and structured environment, simplies the syntax of functions, and also
allows comparisons between dierent versions of the same index, as well as global sensitivity
analysis along the lines of Saisana et al. (2005) (for the distinction between “local” and “global”
sensitivity analysis, see e.g. Saltelli et al. (2019)). COINr also supports time-indexed (panel)
data, represented by the “purse” class (a data frame containing a time-indexed collection of
coins). For more information on coins and purses, see the coins vignette.
All major COINr functions have methods for coins, and many have methods for purses, data
frames, and numerical vectors. This means that COINr can be used either as an integrated
development environment via coins and purses, but equally as a toolbox of functions for other
related purposes.
COINr also oers a far wider range of functions and methodological options than any existing
package. It not only includes a range of options for treating, imputing, normalising and
aggregating indicator data (among others), but also has a suite of analysis tools to check
data availability and perform correlation/multivariate analysis. Moreover, it has many options
for plotting and visualising data using wrapper functions for the ggplot2 package (Wickham,
2016). Many core COINr functions are written with hooks to link with other packages, for
example allowing other imputation or aggregation packages to be used with coins.
Primarily, COINr is used for building composite indicators: In practice this involves assembling
a set of indicators (usually from dierent sources) and accompanying metadata, and assembling
them into data frames that can be read by COINr to build a “coin” (see vignette). After that,
the composite scores are calculated by applying COINr functions to the coin, which specify the
methodological steps to apply, and how to apply them.
To give a avour of COINr, we present a very short example using the built-in “ASEM” data
set which comprises two data frames (one of indicator data, and the other of metadata). To
build a coin, the new_coin() function is called:
# load COINr
# build a coin with example data set
coin <- new_coin(iData = ASEM_iData, iMeta = ASEM_iMeta)
To see how these data frames are formatted, one can use e.g.
View(ASEM_iData) and refer to the coins vignette.
In the most simple case, we could build a composite indicator by normalising the indicators
(bringing them onto a common scale), and aggregating them (using weighted averages to
calculate index scores). This can be done in COINr using the
functions respectively:
# normalise (scale) each indicator onto [0, 100] interval
coin <- qNormalise(coin, dset = ”Raw”,f_n = ”n_minmax”,
f_n_para = list(l_u = c(0,100)))
# aggregate using weighted arithmetic mean
Becker et al. (2022). COINr: An R package for developing composite indicators. Journal of Open Source Software,7(78), 4567. https:
# Note: weights are input in data frames when calling new_coin()
coin <- Aggregate(coin, dset = ”Normalised”,f_ag = ”a_amean”)
Both of these functions allow any other function to be passed to them, allowing more
complex types of normalisation and aggregation. Here, the code simply uses the “min-max”
normalisation method (scaling indicators onto the
[0, 100]
interval), and aggregates using the
weighted arithmetic mean, following the hierarchical structure and weights specied in the
iMeta argument of new_coin().
To see the results in a table form, one can call the get_results() function:
# generate data frame with results at highest aggregation level (index)
get_results(coin, dset = ”Aggregated”)|> head()
uCode Index Rank
1DEU 75.23 1
2GBR 68.94 2
3FRA 65.92 3
4CHE 62.61 4
5NLD 61.24 5
6SWE 60.59 6
We may also visualise the same results using a bar chart - here we see how countries rank on
the “connectivity” sub-index (see Figure 1).
plot_bar(coin, dset = ”Aggregated”,iCode = ”Conn”,stack_children = TRUE)
Figure 1: National connectivity scores broken down into component scores and sorted from highest to
As a nal example, we show one of the analysis features of COINr: the possibility to plot and
analyse correlations.
plot_corr(coin, dset = ”Normalised”,iCodes = list(”Sust”),
grouplev = 2,flagcolours = T, text_colour = ”darkblue”)
Becker et al. (2022). COINr: An R package for developing composite indicators. Journal of Open Source Software,7(78), 4567. https:
Figure 2: Correlations between sustainability indicators, with colouring thresholds. Only correlations
within aggregation groups are shown.
The correlation plot in Figure 2 illustrates where e.g. negative correlations exist within ag-
gregation groups, which may lead to poor representation of indicators in the aggregated
COINr includes far more features than those shown here. Remaining features (with vignette
links) include:
Denomination by other indicators
Screening units by data requirements
Imputation of missing data
Outlier treatment using Winsorisation and nonlinear transformations
Weighting using either manual weighting, PCA weights or correlation-optimised weights.
Analysis via indicator statistics, data availability, correlation analysis and multivariate
analysis (e.g. PCA).
Adjustments and Comparisons: checking the eects of methodological variations.
Global uncertainty and sensitivity analysis of the impacts of uncertainties in weighting
and many methodological choices
A range of visualisation options, including statistical plots, bar charts and correlation
Automatic import from the COIN Tool and fast export to Microsoft Excel.
For the full range of COINr features, see COINr documentation which is accessible at COINr’s
Becker et al. (2022). COINr: An R package for developing composite indicators. Journal of Open Source Software,7(78), 4567. https:
COINr was initally developed under contract for the European Commission’s Joint Research
Centre, and this is gratefully acknowledged for enabling the bulk of the initial design.
Becker, W., Benavente, D., Dominguez Torreiro, M., Tacao Moura, C., Neves, A. R., Saisana,
M., & Vertesy, D. (2019). COIN Tool User Guide. European Commission, Joint Research
Becker, William, Domınguez-Torreiro, M., Neves, A. R., Moura, C. T., & Saisana, M. (2021).
Exploring the link between Asia and Europe connectivity and sustainable development.
Research in Globalization,3, 100045.
Cinelli, M., Spada, M., Kim, W., Zhang, Y., & Burgherr, P. (2021). MCDA index tool: An
interactive software to develop indices and rankings. Environment Systems and Decisions,
41(1), 82–109.
Cobham, A., Janskỳ, P., & Meinzer, M. (2015). The Financial Secrecy Index: Shedding
new light on the geography of secrecy. Economic Geography,91(3), 281–303. https:
Conceição, P. (2020). The 2020 Human Development Report. United Nations Development
Programme. ISBN: 978-92-1-126442-5
Dutta, S., Lanvin, B., & Wunsch-Vincent, S. (2020). The Global Innovation Index 2020:
Who will nance innovation? World Intellectual Property Organisation. https://www.
El Gibari, S., Gómez, T., & Ruiz, F. (2019). Building composite indicators using multicriteria
methods: A review. Journal of Business Economics,89(1), 1–24.
Equal Measures 2030. (2019). Harnessing the power of data for gender equality: Introducing
the 2019 EM2030 SDG gender index. Equal Measures 2030.
Fusco, E., Vidoli, F., & Sahoo, B. K. (2018). Spatial heterogeneity in composite indicator: A
methodological proposal. Omega,77, 1–14.
Lindén, D., Cinelli, M., Spada, M., Becker, W., & Burgherr, P. (2021). Composite Indicator
Analysis and Optimization (CIAO) tool, v.2.
Lindén, David, Cinelli, M., Spada, M., Becker, W., Gasser, P., & Burgherr, P. (2021). A
framework based on statistical analysis and stakeholders’ preferences to inform weighting
in composite indicators. Environmental Modelling & Software, 105208.
R Core Team. (2022). R: A language and environment for statistical computing. R Foundation
for Statistical Computing.
Saisana, M., Saltelli, A., & Tarantola, S. (2005). Uncertainty and sensitivity analysis techniques
as tools for the quality assessment of composite indicators. Journal of the Royal Statistical
Society: Series A (Statistics in Society),168(2), 307–323.
Saltelli, A., Aleksankina, K., Becker, W., Fennell, P., Ferretti, F., Holst, N., Li, S., &
Wu, Q. (2019). Why so many published sensitivity analyses are false: A systematic
Becker et al. (2022). COINr: An R package for developing composite indicators. Journal of Open Source Software,7(78), 4567. https:
review of sensitivity analysis practices. Environmental Modelling & Software,114, 29–39.
Stefana, E., Marciano, F., Rossi, D., Cocca, P., & Tomasoni, G. (2021). Composite indicators
to measure quality of working life in Europe: A systematic review. Social Indicators
Research,157(3), 1047–1078. 021-02688-6
Vraná, L. (2022). Composite Indicators Framework (CIF) for business cycle analysis. Python
Software Foundation.
Wickham, H. (2016). ggplot2: Elegant graphics for data analysis. Springer-Verlag New York.
ISBN: 978-3-319-24277-4
Becker et al. (2022). COINr: An R package for developing composite indicators. Journal of Open Source Software,7(78), 4567. https:
... However, the index framework can be misleading in the absence of transparent indicator development methods (Acosta et al., 2021). Thus, best practices for developing composite indicators from OECD and JRC (2008), Acosta et al. (2019); , and Becker et al. (2017Becker et al. ( , 2022 were used and adapted to normalize and aggregate relevant but diverse and complex indicators into a common unit for measurement and assessment. ...
... While green growth is gaining ground, theoretically and practically, as vital pathways toward sustainable development, its measurement also makes headway. In the last decade, green growth indices, which use composite indices to assess, rank, and compare complex multidimensional concepts that are not immediately measurable, have gained popularity (OECD & JRC, 2008;Becker et al., 2017Becker et al., , 2022. Prominent international organizations developed a plethora of green growth index as sustainable development indicators (Acosta et al., 2019;AfDB, 2014;GGKP, 2013;Jha et al., 2018;OECD, 2017aOECD, , 2023PAGE, 2017;Tamanini & Valenciano, 2016;UNESCAP, 2013;World Bank, 2012). ...
Full-text available
Green growth gained traction as a global climate change strategy and pathway toward sustainable development. China's green growth has been on the rise since the turn of the century, yet it is little understood in the context of its provinces. Previous studies focus on ranking green growth across countries and regions, not on assessing individual provinces over time. This study employs systems thinking and constructs an index framework to assess the environmental, economic, and social dimensions of green growth as a pathway toward sustainable development in Qing-hai on the Qinghai-Tibet Plateau. The study finds that green growth has steadily increased between 2000 and 2021 despite a volatile growth rate. The 10th-13th Five-Year Plans showed similar trends. Short-term green growth performance fluctuated in its dimensions and pillars, while long-term performance increased steadily. Qinghai is well-positioned to achieve sustainable development and build a circular economy. The study further discusses sustainable policy implications.
... In what follows, we report the correlations between indicators in the same subindex, between indicators and their relative aggregates, and finally between dimensions, subindices and the C3 Index. The analysis results and figures presented in this section (Section 3.2) were derived from the COINr package developed by JRC-COIN (Becker et al., 2022). ...
Technical Report
Full-text available
The Cultural and Creative Cities Monitor (hereafter ‘Monitor’) is a monitoring and benchmarking tool first developed in 2017 to allow European cities to compare and contrast their areas of excellence and improvement in terms of culture and creativity. Since its first launch, the Monitor has served as a guidance tool for policy makers at local, national and European level. The 2023 update of the Monitor provides a methodological improvement and a revision of the selected indicators, so to ensure the tool remains reliable and coherent, thus enabling meaningful comparisons of cities’ performance over time. The new results are available for three different reference years, with a focus on the most recent, 2019. These latest results provide a snapshot of the situation just before the COVID-19 pandemic. An updated statistical assessment of the Monitor provided in this report allows the user to use the Monitor consciously, encouraging an informed use of the data and monitoring tools provided. The update has also been applied to the online platform which, on top of the pre-existing tools, also allows users to compare and contrast the performance of 196 European cities at three different points in time.
... The senior liveability index provides a summary assessment of European cities as far as their liveability, and informs broad strategic decisions for the city policy makers. In this paper, we set up our indicator using R 4.2.0 (R Core Team, 2022), and the recently developed COINr package (Becker et al., 2022) for construction and analysis of composite indicators. The latter includes commands for a step-by-step construction of a composite indicator, allowing also for normalisation, weighting, aggregation and the consequent analysis and comparison among multiple configurations. ...
The demographic challenge of the ageing population in European countries needs to be assessed in terms of sustainability and effectiveness of public policies improving the quality of life of the elderly. The European Commission has monitored the quality of life in European cities since 2004, through a survey on citizens’ perception of their liveability. Given that the assessment of life quality embeds a multitude of positive and negative aspects, we develop a composite indicator that we call Senior Liveability Index (SLI) to rank the performance of cities, to monitor the rank changes over time and to explore possible reasons.
Full-text available
Smart mobility systems offers solutions for traffic congestion, transport management, emergency, and road safety. However, the success of smart mobility lies in the availability of intelligent transportation infrastructure. This paper studied smart mobility systems in three Asia-Pacific countries (South Korea, Singapore, and Japan) to highlight the major strategies leading their successful journey to become smart cities for aspiring countries, such as the Kingdom of Saudi Arabia (KSA), to emulate. A robust framework for evaluating smart mobility systems in the three countries and Saudi Arabia was developed based on the indicators derived from the smart mobility ecosystem and three major types of transport services (private, public, and emergency). Sixty indicators of smart mobility systems were identified through a rigorous search of the literature and other secondary sources. Robots, drones, IoT, 5G, hyperloop tunnels, and self-driving technologies formed part of the indicators in those countries. The study reveals that the three Asia-Pacific countries are moving head-to-head in terms of smart mobility development. Saudi Arabia can join these smarter countries through inclusive development, standardization, and policy-driven strategies with clear commitments to public, private, and research collaborations in the development of its smart mobility ecosystem. Moreover, cybersecurity must be taken seriously because most of the smart mobility systems use wireless and IoT technologies, which may be vulnerable to hacking, and thus impact system safety. In addition, the smart mobility system should include data analytics, machine learning, and artificial intelligence in developing and monitoring the evaluation in terms of user experience and future adaptability.
Full-text available
Composite Indicators (CIs, a.k.a. indices) are increasingly used as they can simplify interpretation of results by condensing the information of a plurality of underlying indicators in a single measure. This paper demonstrates that the strength of the correlations between the indicators is directly linked with their capacity to transfer information to the CI. A measure of information transfer from each indicator is proposed along with two weight-optimization methods, which allow the weights to be adjusted to achieve either a targeted or maximized information transfer. The tools presented in this paper are applied to a case study for resilience assessment of energy systems, demonstrating how they can support the tailored development of CIs. These findings enable analysts bridging the statistical properties of the index with the weighting preferences from the stakeholders. They can thus choose a weighting scheme and possibly modify the index while achieving a more consistent (by correlation) index.
Full-text available
In the last two decades, Quality of Working Life (QWL) has become a core element of the European social model and the European Employment Strategy. "More and better jobs" is a strategic goal promoted within Europe for emphasising the attention in QWL. However, there is a large debate in the literature on the definition of QWL, its dimensions, and consequently on the methods to use for its measurement. To the best of our knowledge, the systematic reviews currently available in the literature on QWL measurement in European organisations investigate only a particular industry and/or working population. Moreover, they do not focus specifically on composite indicators, although they appear promising in facilitating QWL understanding and comparisons for supporting decision-makers and policy makers. To overcome these gaps, we conducted a systematic review to identify composite indicators for measuring QWL in European organisations. The review returned 19 studies that are analysed based on a set of factors related to QWL locutions, index name, geographical area, industry or population, level of analysis, dimensions, type of data, inputs, outputs, and test and/or validation. The results highlight a significant heterogeneity among the indicators, confirming the lack of an agreed upon QWL composite indicator for Europe. Such heterogeneity concerns also QWL dimensions. A critical comparison of the different composite indicators is provided, along with a unifying proposal of QWL macro-dimensions. Several gaps in the literature are pointed out suggesting directions for future research.
Full-text available
Asia and Europe have made connectivity between people, businesses and institutions a top political priority in the frame of the Asia-Europe Meeting (ASEM) intergovernmental cooperation forum. In the ASEM context, policy leaders agreed that improving connectivity between countries should contribute to achieve the Sustainable Development Goals. Connectivity is a complex concept involving multiple dimensions. Yet in order to provide clear data-driven evidence to support policymaking on Asia-Europe sustainable connectivity, this concept calls for a measurement framework. The mismatch between the politically adopted definition of sustainable connectivity and existing metrics, led to the development of the two indexes described in this paper. Taken together, they help to explore and understand the relationship between connectivity and sustainability. The results suggest that connectivity and sustainability policy agendas have the potential to mutually reinforce one another. Higher values in connectivity are strongly associated with higher values in the social dimension of sustainability. There are evident gaps between European and Asian nations in terms of connectivity and sustainability. However, both continents are underpinned by the common challenge of reducing the environmental impacts of connectivity without neglecting economic/financial sustainability aspects, while at the same time ensuring that benefits will accrue to society at large.
Full-text available
The Composite Indicator Analysis and Optimization (CIAO) tool v.2 is an expansion of the automated Matlab menu-version of the approach presented by Becker et al. (2017)* for the advanced assessment of Composite Indicators (CIs). The CIAO tool allows users to: 1. Perform a detailed examination of the linear and nonlinear relationships among (i) the set of indicators and (ii) between the indicators and the Composite Indicator (CI). 2. Assess how the indicators are “balanced” within the CI, i.e. to what extent the influence of each indicator matches its assigned weight. 3. Tune the weights of the indicators so that the influence of each one matches pre-defined importance values. *Becker, W., M. Saisana, P. Paruolo, and I. Vandecasteele. 2017. Weights and importance in composite indicators: Closing the gap. Ecological Indicators 80:12-22.
Full-text available
A web-based software, called MCDA Index Tool (, is presented in this paper. It allows developing indices and ranking alternatives, based on multiple combinations of normalization methods and aggregation functions. Given the steadily increasing importance of accounting for multiple preferences of the decision-makers and assessing the robustness of the decision recommendations, this tool is a timely instrument that can be used primarily by non-multiple criteria decision analysis (MCDA) experts to dynamically shape and evaluate their indices. The MCDA Index Tool allows the user to (i) input a dataset directly from spreadsheets with alternatives and indicators performance, (ii) build multiple indices by choosing several normalization methods and aggregation functions, and (iii) visualize and compare the indices’ scores and rankings to assess the robustness of the results. A novel perspective on uncertainty and sensitivity analysis of preference models offers operational solutions to assess the influence of different strategies to develop indices and visualize their results. A case study for the assessment of the energy security and sustainability implications of different global energy scenarios is used to illustrate the application of the MCDA Index Tool. Analysts have now access to an index development tool that supports constructive and dynamic evaluation of the stability of rankings driven by a single score while including multiple decision-makers’ and stakeholders’ preferences.
Technical Report
Full-text available
The COIN Tool is a free Microsoft Excel-based tool designed to help users from research institutions, international organisations, European Union institutions, national and local governments, among others, in the process of building and analysing composite indicators. It was developed by the European Commission's Competence Centre on Composite Indicators and Scoreboards (COIN), at the Joint Research Centre.
Full-text available
Composite indicators are increasingly recognised as a useful tool in policy analysis and public communication. They provide simple comparisons of units that can be used to illustrate the complexity of our dynamic environment in wide-ranging fields, such as competitiveness, governance, environment, press, development, peacefulness, tourism, economy, universities, etc. Their construction has been dealt with from several angles. Some authors claim that MCDM techniques are highly suitable in multidimensional frameworks when aggregating single indicators into a composite one, since this process involves making choices when combining criteria of different natures, and it requires a number of steps in which decisions must be made. In this paper, we conduct a literature review of papers published after 2002 in leading international journals indexed in a recognised database (JCR), in order to identify the different MCDM methods used for aggregating single indicators into composite ones. They have been classified in five categories: the elementary methods, the value and utility based methods, the outranking relation approach, the data envelopment analysis based methods and the distance functions based methods. In general, our review has shown a clear tendency towards an increasing number of papers that use MCDM methods to construct composite indicators since 2014.
Full-text available
Sensitivity analysis provides information on the relative importance of model input parameters and assumptions. It is distinct from uncertainty analysis, which addresses the question ‘How uncertain is the prediction?’ Uncertainty analysis needs to map what a model does when selected input assumptions and parameters are left free to vary over their range of existence, and this is equally true of a sensitivity analysis. Despite this, many uncertainty and sensitivity analyses still explore the input space moving along one-dimensional corridors leaving space of the input factors mostly unexplored. Our extensive systematic literature review shows that many highly cited papers (42% in the present analysis) fail the elementary requirement to properly explore the space of the input factors. The results, while discipline-dependent, point to a worrying lack of standards and recognized good practices. We end by exploring possible reasons for this problem, and suggest some guidelines for proper use of the methods.