The concept of innovation has been defined in various ways, encompassing environmental (Ullah et al., 2021), social (Marini Govigli et al., 2022), technological and institutional innovations (Anwar et al., 2022; Liu et al., 2022). While using a broad definition of innovation can pose challenges in measuring and monitoring, it allows a comprehensive understanding of the concept, which is particularly valuable for interdisciplinary studies.

Several types of innovation have recently emerged as promising drivers for achieving the SDGs, including collaborative innovations (Mariani et al., 2022), eco-friendly innovations (Miao et al., 2023), co-innovation (Adomako & Nguyen, 2022; Fielke et al., 2018) and responsible innovation (Imaz & Eizagirre, 2020; Ranabahu, 2020). The COVID-19 pandemic has increased innovation in, for example, the biopharmaceutical sector, contributing to sustainability (Piñeiro-Chousa et al., 2022). The concept of “innovability” has been proposed by De la Vega and De Paula (2020) as a strategic approach for companies seeking to achieve both competitiveness and sustainability.

To achieve sustainability, significant changes are needed in the public's thinking, behaviour, production and consumption. Scientific expertise alone is not sufficient to develop and implement effective solutions: intellectual capital is also needed, including access to information, knowledge and experience (Faraji et al., 2022). Open innovation principles (Chaurasia et al., 2020), social innovation and disruptive technologies (Ciampi et al., 2020) have all helped to accelerate the sharing of knowledge.

STI distinguishes itself from innovation in its inclusion of the terms “science” and “technology”, thus positioning itself as the driving force behind progress: science can advance knowledge by improving our understanding through observation, experimentation, and research and development (R&D), leading to the creation of techniques and technologies that offer public benefit. Technology is the practical application of scientific knowledge to create products and services. While innovation can result merely from thinking “outside of the box”, technological progress can only be achieved through scientific discovery and research which invent and develop technology (ESCAP, 2015). Technology links science with innovation.

The technology gap is one of the primary causes of persistent inequality between nations (Naudé & Nagler, 2015), and the key difference between developed and developing countries lies in knowledge levels. Eliminating the knowledge gap, as suggested by Ozkaya et al. (2021), can close both development and income gaps. To address this issue, the Addis Ababa Action Agenda acknowledges the role of STI in achieving the SDGs and provides funding to support progress to this end (United Nations, 2015). Furthermore, the Technology Facilitation Mechanism (TFM) and the engagement of United Nations entities and other stakeholders facilitate the exchange of information, best practice, experience and policy advice (United Nations, 2020).

Regarding the frameworks used to understand the SDGs, previous studies have traditionally focused on three principal pillars: social inclusion, economic growth and environmental protection. However, recent global challenges – such as the COVID-19 pandemic, military conflicts and energy and climate shocks – have led to a broader understanding of sustainable development. The 2030 Agenda introduced two additional components, Partnership and Peace, (Dumpe & Guevara, 2020), thus expanding the framework to include five essential dimensions: People (SDGs 1–5), Prosperity (SDGs 7–11), Planet (SDG 6, SDGs 12–15), Peace (SDG 16), and Partnership (SDG 17), known as the 5Ps of the Sustainable Development Goals (Ki-moon, 2019).

To facilitate the monitoring of progress towards the SDGs, an innovative approach was adopted which resulted in a set of indicators for global governance. The Inter-Agency and Expert Group, in collaboration with the Statistical Commission, developed the Global Indicator Framework (GIF) for the SDGs. The GIF comprises 231 unique indicators, some repeated across multiple SDG targets (United Nations, 2022). This approach offers flexibility and adaptability as methodologies and data sources evolve and expand between 2015 and 2030, providing a robust basis for monitoring progress towards the SDGs.

The SDGs do not explicitly define the role that innovation may play in achieving them. However, the specific types of innovation that may drive the achievement of each SDG are not stated too. To better understand the relationship between innovation, STI and the SDGs, we conducted a detailed analysis of the 17 SDG goals and the 169 targets that underpin them, as outlined in the GIF. To this end, we identified the terms linked to innovation and STI in the GIF, first creating a list of all the terms we encountered and labelling them as innovation-related keywords. We were then able to group these keywords into five broad categories: knowledge, science, technology, innovation and information and communication technology (ICT). Finally, we distributed the targets that included the keywords into these categories, which we then aligned with the 5Ps, as shown in Table 1, which is formed of five columns representing the five categories (Knowledge, Science, Technology, Innovation and ICT) and five rows representing the 5Ps (People, Prosperity, Planet, Partnership and Peace). We thus identified the targets or terms that we expected to find in the sample database. From this basis, it was then possible to compare the expected and actual occurrences of innovation-related keywords in the literature by examining the terms, and the frequency with which they appeared, using the SDG mapping tool.

Table 1.

Distribution of innovation-related SDG goals and targets by 5Ps.

	Knowledge	Science	Technology	Innovation	ICT
People	SDG 2.3		SDG 1.4		SDG 4.4
		SDG 2.a agricultural research, technology			SDG 4.b
		SDG 3.b medical research	SDG 6.a recycling, reuse technology		SDG 5.b
Prosperity			SDG 7.1 SDG 7.b	SDG 9	SDG 8.10
		SDG 7.a clean energy research technology		SDG 8.2	SDG 9.c
			SDG 9.2 SDG 9.4 SDG 9.a	SDG 8.3
		SDG 9.5/9.b scientific research, technology, innovation
Planet		SDG 12.a scientific and technological capacity
		SDG 14.3	SDG 12.5
		SDG 14.4
	SDG 14.a scientific knowledge, research, technology
Partnership			SDG 17.7
	SDG 17.6 STI and enhance knowledge-sharing through a global technology facilitation mechanism
	SDG 17.8 technology bank and science, technology and innovation capacity-building mechanism, ICT
	SDG 17.16 share knowledge, expertise, technology
Peace	SDG 16.10

Source: Prepared by the authors based on screening of the Global Indicator Framework

We identified 31 (out of 169) targets that mentioned at least one innovation-related term. The term “innovation” was itself mentioned in SDG 9 and targets 8.2, 8.3, 9.5 and 9.b. The term “STI” was observed in SDGs 17.6 and 17.8, and the terms “research” and “science” were repeatedly mentioned, with 14 and 9 occurrences respectively. However, the most frequently used term was “technology” (including ICT), which appeared 33 times. Technology was found to be relevant across four dimensions (People, Prosperity, Planet and Partnership). However, it is important to note that technology alone cannot achieve sustainable development, due to the possibility of rebound effects (Giovannini & Roure, 2017). The United Nations recognises the crucial role played by technology but, at the same time, highlights the need for it to be integrated with science and innovation to achieve the SDGs (Sachs et al., 2022).

To ensure the reproducibility of our results, we established precise inclusion and exclusion criteria and performed a comprehensive search of keywords, titles and abstracts across both databases. Our preliminary review revealed that authors frequently mention the SDGs and innovation in their abstracts to appeal to a wider audience. To narrow our search, we focused particularly on article titles, as our preliminary review showed that authors generally include the term “innovation*” in their title if innovation is a core aspect of their study. We first used the search query: title, abstract, keywords = (“sdg*” OR “sustainable development goal*”) AND title = “innovation*”. This query identified 1374 records (Scopus n = 826; WOS n = 548) (Fig. 1). We then conducted a second query: title, abstract, keywords = “sdg*” OR “sustainable development goal*” AND “science, technology, and innovation”, which yielded 91 records (Scopus n = 60; WOS n = 31). Overall, our initial search generated 1465 results.

Fig. 1.

PRISMA flow diagram and methods.

(0.43MB).

Sources: Figure created by the authors

We applied the following filters: (a) Document type: Articles, (b) Language: English, (c) Publication years: 2015–2023, and cleaned the data to remove duplicate entries and address any missed keywords, titles or abstracts. Following these procedures, 571 relevant records remained.

We applied further exclusion criteria – (a) SLRs and (b) bibliometric analyses – to narrow the selection during the screening process. Following a content analysis of titles and abstracts, 15 SLRs and 12 bibliometric analyses were excluded, resulting in a total of 544 studies remaining for further analysis (Fig. 1). From this sample, it can be seen that the number of publications has grown exponentially over the period in question, from only two in 2015 to 185 as of 11 November 2023.

The analysis in this study was performed using VOSviewer software version 1.6.19, a tool designed to create and visualise co-occurrence networks. We utilised two types of visualisation: network visualisation and cluster density visualisation; both used only the keywords for each article in our sample – those chosen by the authors and those generated automatically by the Keyword Tool.

In the network visualisation, keywords are depicted as labels accompanied by circles, where the size of the circle corresponds to the weight of the keyword. The weight of a keyword reflects its importance and is calculated through various metrics, including link weight, total link strength weight and occurrence weight. The score attribute indicates any numerical properties of a keyword (e.g. average publication year score or average citation score). The VOSviewer technique groups closely linked keywords into individual clusters which are colour-coded. The colour of a circle is thus determined by the cluster to which it belongs, and links between circles are represented as lines (van Eck & Waltman, 2023).

Cluster density visualisation, in contrast, focuses on the occurrence and popularity of keywords. Similarly to network visualisation, keywords are represented as points, and the saturation of the point's colour is determined by the number of keywords in its neighbourhood and the weight of the neighbouring keywords.

Both visualisations offer a comprehensive and visually appealing representation of networks and clusters, enabling a deeper understanding of the relationships between the terms within the sample database and their relative densities (van Eck & Waltman, 2023).

To answer RQ2, we utilised a web platform called “Knowledge Base for the Sustainable Development Goals”, which includes the SDG Mapper tool. This platform serves as a knowledge hub for EU policies, indicators, and data related to the SDGs. The SDG Mapper uses natural language processing (NLP) techniques to detect the SDGs mentioned in the uploaded bibliometric records (European Commission, 2021). Once references to the SDGs have been identified, machine learning algorithms are then applied to analyse the data in the sample database and establish connections between the bibliometric records and the relevant SDGs. This process utilises rule-based techniques to identify specific keywords or phrases associated with each SDG.

The results of the analysis are presented in various visualisations. The bar charts provide an overview of the SDGs identified and their relative importance in the text, offering insights into the frequency and prominence of each SDG. The bubble charts show the relevance of the SDG goals and targets within the records, illustrating the distribution and interconnectedness of the SDGs. These visualisations enable an understanding of the representation and importance of different SDGs within the analysed bibliometric records, and facilitate a comprehensive mapping and analysis of SDG-related knowledge.

The text-mining feature of the VOSviewer facilitated the generation of a term map that relies on NLP algorithms (van Eck & Waltman, 2011). In contrast to the network and cluster density visualisations, the text-mining technique examines the titles and abstracts of the sample database. In the VOSviewer, a noun phrase is defined as a sequence of one or more consecutive words in which the last word is a noun and any preceding words are nouns or adjectives. The two-dimensional map shows the placement of terms according to their relatedness, with closer proximity indicating a stronger relationship. This tool determines the relatedness of terms based on their co-occurrence within the records.

A text file was prepared to identify the relevant SDG goals and targets within the sample database of 544 bibliometric records, including titles, abstracts and keywords. The SDG Mapper detected 122 SDG targets (out of 169) in the sample database.

Table 3 shows the 24 targets with innovation-related keywords that were identified by the SDG Mapper tool. When we compared these to our original list of 31 targets that we expected to find within the sample database, we found 18 matches (marked in bold). An examination of the sum of occurrences shows that researchers are most interested in SDG 8.2 (n = 514 occurrences), SDG 9.5 (n = 452), and SDG 7.2 (n = 353).

The relevance of goals and targets is represented by bubbles in Fig. 4. The size of the bubble is determined by the percentage of corresponding keywords detected (the ratio of the number of keywords relating to a single goal to the total number of keywords). The total number of detected keywords was 3656. Thus, SDG 8 (decent work and economic growth) is the most visible, with 1107 occurrences, followed in second and third places by SDG 9 (industry, innovation and infrastructure) and SDG 7 (affordable and clean energy), with 567 and 534 occurrences respectively.

Fig. 4.

SDG targets based on the percentage of corresponding detected keywords.

(0.38MB).

Sources: Figure created by the authors using SDG Mapper tool

SDG 8.2 (diversify, innovate and upgrade for economic productivity) is represented by the largest bubble, with “technolog innov” the most detected keyword, appearing 494 times in the sample database. The second most popular keyword was “renew energi”, with 282 occurrences, followed by “econom growth”, which appeared 167 times. The top ten keywords also include “green technolog” (n = 122 occurrences), “sti” (n = 60) and “r&d” (n = 59).

Fig. 5 shows the SDGs detected and their distribution across the 5Ps. The consolidation of the SDGs into the 5Ps supports efforts to gauge progress and underscores the fact that the SDGs are not a collection of isolated goals but, rather, closely interconnected.

Fig. 5.

. Relevant SDG goals according to 5Ps.

(0.27MB).

Sources: Figure created by the authors using SDG Mapper tool

The results of the SDG Mapper show that the Prosperity dimension, which includes SDG 7 (14.6 %), SDG 8 (30.3 %), SDG 9 (15.5 %), SDG 10 (reduced inequalities) (3.1 %) and SDG 11 (sustainable cities and economies) (3.1 %), accounts for approximately 67 % of all identified keywords. Thus, the majority of the publications in our database were devoted to economic aspects. The next most frequently occurring SDGs are those pertaining to the Planet dimension, including the environmental aspects of SDG 6 (clean water and sanitation) (1.1 %), SDG 12 (responsible consumption and production)(5.1 %), SDG 13 (climate action) (7.1 %), SDG 14 (life below water) (0.2 %) and SDG 15 (life on land) (5.3 %) which collectively account for 18.8 % of occurrences. The People dimension (11.8 %) ranks third, encompassing SDG 1 (no poverty) (1.8 %), SDG 2 (zero hunger) (3.1 %), SDG 3 (good health and well-being) (4.9 %), SDG 4 (quality education) (1.6 %) and SDG 5 (gender equality) (0.4 %), which together address equity and justice within public and private communities and national entities. SDG 17 (partnership for the goals), representing the Partnership dimension, is crucial in fostering innovation but was penultimate in the list of most frequently cited SDGs, with 1.9 %. The Peace dimension (SDG 16 – peace, justice and strong institutions) is mentioned in the fewest publications, accounting for only 0.9 % of mentions.

The text-mining technique identified concepts, that is, the most frequently appearing noun phrases in titles and abstracts. Noun phrases that appeared at least ten times were, in accordance with the default position, considered to meet the threshold for inclusion. Of the set of 13,163 noun phrases, 439 met this threshold. A crucial metric in text mining is the relevance score, which ranks words based on their importance in abstracts and titles. As might be expected, the most commonly occurring keywords were “sdgs” (n = 919 occurrences) and “innovation” (n = 892), but these were excluded as they formed part of the initial search query. Moreover, these terms had relevance scores of 0.02 and 0.21 respectively and, as argued by Van Eck and Altman (2023), excluding terms with low relevance scores – which are often generic terms and contribute minimal information – enhances the usefulness of a map. With these two terms hidden, the mapping was then rescored, resulting in 437 frequently occurring noun phrases which were selected for display on the map.

The VOSviewer text-mining network formed five clusters of varying sizes. The most prominent was a red cluster containing 200 items, while the smallest was a purple cluster with only 10 items.

Fig. 6 presents a network of distinct clusters of interconnected concepts precisely identified through co-occurrence frequencies. Each cluster, depicted in a different colour, showcases a different dimension of the role of innovation as discussed in the literature. We labelled the red cluster “Promoting Sustainable Development”: it involves addressing issues in a way that supports the planet, benefits society and ensures long-term economic viability. This cluster highlights the interplay between government, universities and industry. The innovation-related keywords in this cluster are similar to those in the network map (Table 2).

Fig. 6.

Text-mining visualisation labelled by general topics.

(0.83MB).

Sources: Figure created by the authors using VOSviewer 1.6.19

The green cluster, “Driving Economic Growth”, investigates “impact” on “economic growth” and the use of “renewable energy” and “technological innovation” to achieve environmental sustainability by reducing “CO2 emissions”. This cluster emphasises leveraging “green innovation”, “eco-innovation” and “environmental innovation” to drive economic growth while reducing environmental impact and addressing climate change. When supported by financial innovation, aspects such as “renewable energy” and “energy innovation” play a significant role in ensuring long-term sustainability.

The blue cluster, “Strengthening policy”, is concerned with policies and initiatives that promote innovation and sustainability within specific geographic regions. The cluster also includes “methodology”, “index” and “assessment”, which refer to the development and adoption of the Agenda 2030 programme. The yellow cluster, “Enhancing enterprise performance”, focuses on systematically examining data and evidence and utilising “models” and “theories” to understand “industry” and “firm performance”, while the purple cluster brings together various keywords that were not grouped with other clusters.

Previous studies have explored the role of innovation in various contexts and specific types of innovation (Anwar et al., 2022; Khan et al., 2022; Schroeder et al., 2019; Wang et al., 2022). We aimed to add value to the existing literature by examining the broader contribution of innovations in achieving the SDGs.

First, by explicitly focusing on three keywords – “innovation”, “sti” and “sdgs” – we were able to gain a comprehensive overview of the intellectual structure of the existing literature, which we found to be organised into thematic clusters around the roles of innovation in achieving the SDGs.

Second, our study highlights the explicit connections between innovation and specific SDGs. Using the SDG Mapper tool, we identified 1742 keywords, showing the domains where innovation is closely intertwined with the SDGs. However, the relationship between innovation and the SDGs remains underexplored in some significant areas, including gender equality (SDG 5), peace (SDG 16) and partnership (SDG 17), presenting many opportunities for further research. This valuable information may guide future research endeavours and policy-making initiatives, directing attention towards specific SDGs and target areas where innovation can have the most substantial influence.

Thirdly, our study aimed to identify research gaps and propose potential avenues for future investigation. By emphasising the need for more comprehensive literature on STI and identifying research gaps related to the partnership dimension and TFM, we highlight valuable directions for researchers to explore in the future. These insights not only contribute to the existing knowledge base on innovation and the SDGs but also encourage further enquiry into critical areas that warrant attention, enabling a deeper comprehension of the available studies and their interconnectedness. Our findings suggest that scholars should adopt a broad perspective when approaching innovation, as it intersects with various research domains that play a crucial role in fostering a holistic understanding and offer the potential for engaging studies.

It is essential to acknowledge that our analysis has certain limitations: the first relates to the choice of keywords. While we used a comprehensive set of keywords, the possibility remains that some relevant studies were omitted. Expanding the set of keywords or using alternative search strategies could help capture a broader range of literature on innovations and the SDGs.

The second limitation pertains to the use of specific databases and exclusion criteria. By focusing on the Scopus and WOS databases and excluding certain document types, relevant studies may have been missed.

The third limitation relates to the methodology employed in our study. While we used three interrelated methods (co-occurrence analysis, SDG mapping and text-mining), alternative methodologies exploring different approaches, such as qualitative content analysis, may provide additional insights and perspectives.