Digital Pathways Toward Sustainability in Agri-Food Waste: A Systematic Review

Abstract

The growing environmental and economic impacts of agri-food waste have intensified interest in digital and circular strategies for more sustainable resource management. This study investigates how digital technologies are being applied to enhance the circular management of agri-food waste. A systematic literature review was conducted following PRISMA guidelines, using Scopus as the primary database. The bibliometric analysis included 373 publications from 2015 to 2025 and was complemented by a thematic review of the 20 most cited articles. Results revealed six major research clusters, ranging from predictive waste analytics and smart traceability systems to circular bioeconomy applications such as anaerobic digestion and pyrolysis. In addition, the study examined structural indicators such as food waste per capita and hunger indices to contextualize geographic disparities in research output. The findings underscore the dual role of digital technologies as both operational tools and mechanisms for reducing systemic inequalities. Overall, the integration of intelligent systems and circular models offers promising pathways to support the Sustainable Development Goals and foster a more inclusive and resilient agri-food sector.

1. Introduction

The sustainable management of resources is one of the most pressing challenges facing both economic and environmental systems today, particularly in the agri-food sector, where waste along the supply chain continues to have a significant ecological, social, and economic impact. This calls for a paradigm shift: transforming waste into a new productive resource [1]. Each year, millions of tons of food are lost or wasted, contributing not only to the inefficient use of natural resources but also to increased greenhouse gas emissions associated with the treatment and disposal of organic waste [2,3].

Table 1. Food waste and CO₂e emissions in the European Union (2018–2022). Author’s elaboration.

Year	Estimated Food Waste (Million Tons)	Estimated Disposal-Related Emissions (Million Tons CO2e)	Kg CO2e per kg of Food Waste	Source
2022	59	~186	~3.15 kg CO2e/kg	[4]
2021	~58 (estimate)	~183	~3.15 kg CO2e/kg	[5]
2020	—	—	—	—
2019	~58–60	~180–190	~3.10–3.20 kg CO2e/kg	[5]
2018	~59	~185	~3.15 kg CO2e/kg	[6]

quantifies this relationship by presenting data for the period 2018–2022 in Europe, highlighting the correlation between food waste and CO₂e emissions.

In the context of ecological and digital transitions, adopting circular economy models supported by digital technologies has emerged as a strategic priority. These models aim to reduce waste and valorize agri-food by-products, avoiding excessive consumption of resources, as discussed by Bahn et al. [7], who analyze the potential and risks of digitalization for sustainable agri-food systems. The United Nations’ 2030 Agenda offers a policy framework through key Sustainable Development Goals such as SDG 9 (Industry, Innovation, and Infrastructure), SDG 12 (Responsible Consumption and Production), and SDG 13 (Climate Action), guiding both public and private efforts in this direction. In particular, Cingolani [8] highlights the role of financial instruments and governance mechanisms in aligning public and private investment strategies with these goals.

The European Green Deal and the “Fit for 55” package outline a clear roadmap for more efficient management of material and energy flows. Within this framework, the agri-food sector faces a dual transition: on one hand, toward reducing waste and recovering organic waste as a new raw material; on the other, toward integrating advanced technological solutions that enable intelligent systems for monitoring, forecasting, and resource valorization [9,10].

Digital technologies are rapidly emerging as key enablers of more effective and sustainable food waste management. They allow real-time data collection and analysis across the supply chain, product traceability, logistical optimization, and the activation of collaborative redistribution models [11,12]. At the same time, predictive tools and intelligent algorithms help anticipate surpluses, minimize losses, and support more informed decision-making at both industrial and household levels, ultimately safeguarding natural resource use [13].

However, increasing digitalization also brings new challenges, including system interoperability, equitable access to technology, and the protection of sensitive data [14,15]. In this scenario, policy plays a crucial role in shaping a regulatory environment conducive to the large-scale adoption of digital solutions, ensuring inclusivity, security, and sustainability [15].

This article critically examines how digital technologies support the shift toward circular agri-food waste management models. Building on these premises, the study provides a systematic and up-to-date analysis of the current state of the art in the scientific literature. The structure of the work guides the reader through a coherent and methodologically transparent path. The approach is based on a bibliometric analysis of key contributions indexed in the Scopus database [16,17], followed by a thematic evaluation of the 20 most cited articles [18,19], with the goal of identifying core conceptual clusters, emerging research directions, and future technological developments in the field.

Despite increasing attention to the role of digital technologies in supporting circular economy strategies, the literature remains fragmented, with few studies offering a structured and quantitative overview of this emerging research domain. This study addresses this gap by combining bibliometric analysis with qualitative thematic coding, offering a comprehensive mapping of digital enablers for food waste valorization.

2. Materials and Methods

The methodology adopted, in alignment with the approach proposed by Campana et al. (2025) [16,17], consists of two main phases: a systematic literature selection and a bibliometric analysis.

The first phase was carried out following the PRISMA guidelines, with the aim of identifying the most relevant academic contributions on the topic of sustainable agri-food waste management, particularly in relation to the use of digital technologies. The Scopus database was used as the primary source, applying a structured set of inclusion and exclusion criteria to ensure thematic consistency and scientific rigor in the selection of publications [18,19]. Scopus was chosen due to its status as one of the most comprehensive and widely recognized bibliographic resources, offering high-quality academic content and broad disciplinary coverage. For these reasons, it has become a reference database for conducting bibliometric analyses and systematic literature reviews [17].

The second phase involved a bibliometric analysis using the VOSviewer software, version 1.6.20, aimed at mapping the main research trajectories in the field. Specifically, emerging themes, keyword co-occurrences, and collaboration networks among authors were identified [20].

The integration of the PRISMA approach with bibliometric analysis provided a systematic and up-to-date overview of the evolution of the scientific literature on the application of digital technologies in agri-food waste management. In this context, an in-depth review of the 20 most cited articles in the corpus was also conducted, with the goal of highlighting the most influential contributions, prevailing theoretical models, and key future research directions.

2.1. Systematic Literature Selection and Inclusion Criteria

To ensure a systematic and reproducible analysis of the scientific literature on the application of digital technologies in the sustainable management of agri-food waste, a methodology inspired by the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework was adopted [21,22,23].

The literature search was conducted using the Scopus database, through an advanced query combining key terms and relevant synonyms from the following conceptual domains: (i) agri-food waste, (ii) resource management, valorization, and recovery, (iii) environmental sustainability, and (iv) digital technologies. The search string was constructed using Boolean operators (AND, OR) to balance breadth and thematic relevance. The full query used is as follows:

(TITLE-ABS-KEY(“agri-food waste” OR “food waste” OR “agricultural waste” OR “agro-industrial waste” OR “biowaste” OR “organic waste” OR “food processing waste” OR “agri-food by-product*” OR “agricultural by-product*” OR “food supply chain waste” OR “food loss” OR “surplus food” OR “expired food” OR “discarded food” OR “waste biomass” OR “residual biomass”))

AND (TITLE-ABS-KEY(recycl* OR reus* OR “valorization” OR “valorisation” OR “upcycling” OR “resource recovery” OR “circular economy” OR “waste management” OR “bioeconomy” OR “waste to energy” OR “waste to resource” OR “biorefinery” OR “composting” OR “anaerobic digestion” OR “bioconversion”))

AND (TITLE-ABS-KEY(sustainab* OR “green technology” OR “clean technology” OR “eco-innovation” OR “environmental impact” OR “resource efficiency” OR “Agenda 2030” OR “sustainable development goal*” OR SDG* OR “climate impact” OR “environmental sustainability” OR “low carbon” OR “carbon footprint” OR “life cycle assessment” OR LCA))

AND (TITLE-ABS-KEY(digital* OR “industry 4.0” OR “digital transformation” OR “smart agriculture” OR “precision agriculture” OR “IoT” OR “AI” OR “artificial intelligence” OR “machine learning” OR “blockchain” OR “data analytics” OR “remote sensing” OR “big data” OR “ICT” OR “sensor-based system*” OR “automation” OR “cyber-physical system*”))

The initial execution of the query, without any filters, yielded 524 documents. A temporal filter was then applied, limiting the selection to the period from 2015 to 2025 to ensure the relevance and timeliness of the analyzed contributions. This reduced the dataset to 513 documents.

To ensure scientific quality, a document type filter was applied, including only primary academic sources, i.e., peer-reviewed journal articles and conference proceedings [24,25,26]. Following this screening phase, the dataset was narrowed to 436 documents.

A disciplinary filter was subsequently applied, including only publications classified under the following subject areas: Environmental Science, Energy, Computer Science, Agricultural and Biological Sciences, and Business, Management, and Accounting, deemed most relevant to the objectives of this study. After this selection, the final dataset consisted of 373 documents. Figure 1 presents the document selection pathway according to the PRISMA framework, outlining each step from initial identification to final inclusion.

The inclusion and exclusion criteria applied at each phase are summarized in

Table 2. Inclusion and exclusion criteria applied during the document selection process across the four phases of the PRISMA framework. The screening process focused on relevance, quality, and disciplinary alignment. Author’s elaboration.

Phase	Inclusion Criteria	Exclusion Criteria
Identification	Documents indexed in Scopus between 2015 and 2025	Documents outside the selected period (pre-2015) (n = 11)
Screening	Peer-reviewed journal articles and conference proceedings	Notes, editorials, book chapters, theses, and other non-peer-reviewed materials (n = 77)
Eligibility	Belonging to the following disciplines: Environmental Science, Energy, Computer Science, Agricultural and Biological Sciences, Business	Publications in unrelated fields (e.g., general social sciences) (n = 63)
Inclusion	Documents meeting all previous criteria and aligned with the thematic scope of the research	(Final records included: 373)

.

During the identification phase, records were retrieved through the application of the defined Boolean string. The screening phase involved the exclusion of materials not subjected to peer review. The eligibility phase included a manual assessment of each document’s disciplinary relevance. Finally, in the inclusion phase, only those contributions that fully met both the methodological and thematic criteria were retained for analysis [27,28].

2.2. Bibliometric Analysis

From a bibliometric perspective, VOSviewer is one of the most widely used tools for the graphical representation and analysis of scientific literature networks. It enables the structured visualization of information such as titles, authors, keywords, and citations [29,30]. Due to its versatility, the software facilitates the identification of research trends, areas of impact, and evolutionary trajectories through the analysis of co-occurrences among citations and keywords [31].

Within the academic landscape, VOSviewer has established itself as an essential tool for the visual analysis of bibliometric data, aiding in the identification of knowledge gaps, thematic concentrations, and key sources in specific disciplinary domains [32,33]. The software primarily operates at an aggregated level, making it particularly effective for thematic cluster analysis [32].

In this study, the clustering algorithm embedded in version 1.6.20 of VOSviewer was applied, allowing the aggregation of recurrent keywords into distinct thematic groups, each identified by a specific color [34]. In these graphical representations, the size of the nodes (circles) indicates the frequency of keyword usage, while colors distinguish the different clusters. The spatial coordinates of the nodes along the x and y axes do not carry specific semantic meaning, and the visual layout can be rotated or mirrored without affecting the underlying conceptual relationships [35,36].

The dataset used for the analysis was exported in .CSV format from the Scopus database, ensuring full compatibility with VOSviewer and proper processing of bibliographic metadata. The bibliometric map was generated based on both author-provided and indexed keywords. To improve readability and visualization consistency, a threshold filter was applied to include only keywords that appeared at least three times in the corpus.

This procedure enabled the clear identification of the most relevant concepts and areas of highest semantic density within the research domain, producing an effective map of emerging terms and major thematic clusters in the scientific literature on the sustainable management of agri-food waste through digital technologies [37].

2.3. Qualitative Thematic Analysis of Top Cited Articles

In addition to the bibliometric mapping, a qualitative thematic analysis was conducted on the 20 most cited articles within the selected corpus [38]. Each contribution was thoroughly reviewed and coded using an inductive approach, allowing key themes to emerge directly from the content rather than from a predefined taxonomy.

The methodological framework draws on the principles of qualitative interpretive analysis grounded in grounded theory [39,40], enabling the identification of the main research directions present in the literature. The themes that emerged were then cross-referenced with the clusters identified in the bibliometric analysis, in order to ensure conceptual coherence, thematic saturation, and interpretive robustness.

To select the 20 most cited articles, citation count within the final Scopus-derived dataset (n = 373) was used as the primary criterion, based on the assumption that high-impact publications are indicative of influential research in the field [41]. However, a thematic screening was also conducted to ensure that the selected articles were strongly aligned with the research objectives, specifically, the intersection of digital technologies, agri-food waste, and circular economy. Highly cited works that were marginal to this scope were excluded. It is acknowledged that citation-based selection may introduce certain biases, particularly in favor of open-access journals, which tend to receive higher visibility and broader citation counts due to accessibility advantages [42]. To mitigate this, the analysis included publications from a diverse range of disciplines and journals indexed in Scopus, ensuring thematic relevance across Environmental Science, Agricultural and Biological Sciences, Energy, Computer Science, and Business and Management domains.

3. Results

3.1. Bibliometric Results: Keyword Co-Occurrence and Cluster Mapping

The bibliometric analysis performed through VOSviewer produced the network visualization presented in Figure 2.

The analysis of the keyword co-occurrence map, generated using VOSviewer software based on the selected dataset, clearly visualizes the main emerging thematic trajectories in the scientific literature related to the sustainable management of agri-food waste within the framework of the digital circular economy. Each node in the network represents a recurring keyword, while the links between nodes indicate the frequency of co-citations, reflecting the semantic density and conceptual proximity between the addressed topics. The different clusters, color-coded, highlight areas of greater concentration and theoretical interconnection, allowing exploration of the internal structure of the research field.

Specifically, the red cluster, located on the left side of the map, focuses primarily on the social, behavioral, and managerial aspects related to food waste and municipal solid waste management. Keywords such as waste management, municipal solid waste, food loss, cost, age, and community influence indicate a widespread focus on household and community dynamics of food waste, as well as on the socio-demographic factors that influence its generation. This cluster reflects the interest in integrating traditional management tools with predictive models applied in urban contexts, in line with literature contributions analyzing collection infrastructure resilience, consumer behavior, and multilevel governance of food surplus.

The orange cluster, strongly interconnected with the red one, is centered around the keyword machine learning and represents the growing relevance of intelligent technologies in waste prediction and management. Within it appear terms such as algorithm, controlled study, and feedforward neural network, highlighting the use of advanced computational models for data analysis, waste flow forecasting, and the optimization of operational decisions. This area aligns with the expansion of artificial intelligence technologies, widely discussed in the article, as a driver of the digital transformation of agri-food supply chains and the achievement of SDG 9 and SDG 12.

At the center of the map lies the purple cluster, which groups keywords related to the circular economy (circular economy, enzymatic conversion, microbial fermentation, and 3D food printing). This group represents a convergence between technological and systemic approaches, with a focus on innovation in biochemical processes and the reconversion of waste into value-added resources. The emergence of these topics reflects the literature on the use of by-products for the production of new materials or ingredients, such as alternative proteins, as well as attention to regenerative models inspired by the bioeconomy and industrial symbiosis.

The green cluster, located on the right side of the visualization, gathers keywords such as biomass, bio-oil, microwave pyrolysis, and integrated energy system, linked to energy valorization technologies for organic waste. References to processes such as pyrolysis and anaerobic digestion reveal a strong orientation toward the transformation of agri-food waste into renewable energy carriers, in line with the principles of the circular bioeconomy. This research stream intersects with the growing interest in environmental impact assessment methodologies, such as life cycle assessment, and in the contribution of these technologies to the decarbonization of production systems (SDG 13).

Finally, the blue cluster, located at the far right of the map, synthesizes the role of artificial intelligence and emerging digital technologies in supporting integrated circular economy strategies. Keywords such as biofuels, bioenergy, and digital twin draw attention to advanced digital solutions, such as virtual modeling, predictive simulation, and automated traceability, applied to the sustainable management of bioenergy resources. This group highlights how the integration of AI, IoT, and cyber-physical systems can act as a catalyst for circular innovation in agri-food systems, contributing to increased efficiency and resilience of supply chains.

Table 3. Top 10 keywords by Total Link Strength in the co-occurrence network. Authors’ elaboration.

Rank	Keywords	TLS
1	machine learning	118
2	circular economy	101
3	controlled study	100
4	pyrolysis	87
5	anaerobic digestion	78
6	biomass	74
7	optimization	74
8	review	74
9	article	65
10	nonhuman	59

Note: TLS = Total Link Strength.

presents the top 10 most frequently co-occurring keywords identified in the bibliometric analysis, illustrating the underlying conceptual architecture depicted in the visualization. At the top of the ranking is machine learning (TLS = 118), underscoring the centrality of algorithmic and data-driven approaches in the current scientific discourse on sustainable waste management. This keyword reflects a robust emphasis on predictive modeling and artificial intelligence methods applied to the optimization of resource flows and waste reduction strategies.

Circular economy (TLS = 101) follows closely, signaling the strategic alignment of digital innovation with systemic sustainability principles. This concept bridges multiple thematic areas, reflecting the cross-cutting role of circularity in both technical and policy-oriented research. Controlled study (TLS = 100) and optimization (TLS = 74) reinforce the prominence of experimental, model-based, and performance-driven approaches, highlighting the shift toward evidence-based frameworks in the design and evaluation of waste management interventions. The presence of pyrolysis (TLS = 87), anaerobic digestion (TLS = 78), and biomass (TLS = 74) indicates a strong focus on biochemical and thermochemical conversion technologies, revealing a parallel research stream concerned with energy recovery and valorization of organic waste, a trend particularly salient within the green cluster. Lower in the ranking but still significant are generalist terms such as review (TLS = 74) and article (TLS = 65), which reflect the growing consolidation of this field and the emergence of structured syntheses of existing knowledge. The inclusion of nonhuman (TLS = 59) points to the increasing role of machine-led and sensor-based systems in automation and environmental monitoring processes.

Together, these keywords delineate a highly interconnected and multidisciplinary research landscape. The interplay between computational modeling, resource recovery, and systemic sustainability frameworks suggests that the domain of agri-food waste management is rapidly evolving through the convergence of digital technologies, bio-based innovation, and policy integration.

The interconnections among the most relevant keywords highlight how sustainable waste management is increasingly being approached through a multidimensional lens, integrating computational intelligence, circular economy principles, and technological pathways for resource valorization. The centrality of terms such as machine learning, controlled study, and optimization reflects the growing adoption of artificial intelligence-based methodologies and predictive models that support more informed decision-making and enhanced operational efficiency. The recurrence of keywords such as pyrolysis, anaerobic digestion, and biomass indicates a strong interest in biochemical and thermochemical conversion technologies, which enable the recovery of energy and materials from organic waste, in alignment with the core principles of the circular economy. The prominent ranking of the keyword circular economy further confirms the strategic relevance of systemic integration between technological innovation and environmental sustainability.

3.2. Journals and Country-Level Research Distribution

Table 4. Leading journals contributing to the research domain, listed according to the volume of relevant publications. Our elaboration.

Rank	JT	TNP
1	Sustainability	22
2	Waste Management	17
3	Science of the Total Environment	16
4	Journal of Cleaner Production	15
5	Journal of Environmental Management	13
6	Bioresource Technology	12
7	Waste Management and Research	7
8	Lecture Notes in Civil Engineering	7
9	IFIP Advances in Information and Communication Technology	7
10	Trends in Food Science and Technology	6

Note: JT = Journal title; TNP = Total No. of publications.

presents the top 10 scientific journals by number of publications related to the topic of sustainable agri-food waste management supported by digital technologies, with particular attention to artificial intelligence, digital twins, and the circular economy. Leading the ranking is the journal Sustainability, with 22 publications, followed by Waste Management (17) and Science of the Total Environment (16). These are all key journals for exploring interdisciplinary approaches to waste reduction, advanced recycling, and the implementation of predictive models.

Other noteworthy journals include the Journal of Cleaner Production (15) and the Journal of Environmental Management (13), which remain reference points for methodological innovation in environmental sustainability and integrated waste management. The presence of journals such as Bioresource Technology (12) and Waste Management and Research (7) further highlights the strong interest in organic waste valorization technologies and bioenergy practices.

Of particular interest is the inclusion of outlets such as Lecture Notes in Civil Engineering and IFIP Advances in Information and Communication Technology, which demonstrate the growing role of digital technologies and smart solutions in traditionally engineering- and ICT-oriented fields. Trends in Food Science and Technology closes the ranking with six contributions, confirming the increasing attention to food waste and its technology-driven management within the agri-food supply chain.

Overall, this editorial distribution reflects the growing scientific visibility and relevance of data-driven solutions for sustainable development, offering researchers and practitioners a set of essential reference points to explore emerging trends, case studies, and methodological advances in the field of intelligent resource and waste management.

Table 5. Top 10 countries by number of publications on digital and circular approaches in agri-food waste management. Our elaboration.

Country	TNPs	TNCs	TLS
China	39	149	23
Egypt	22	139	18
Germany	25	167	14
Nigeria	16	52	14
Portugal	14	64	12
Spain	9	17	10
Hungary	8	65	8
Saudi Arabia	3	90	7
India	2	21	6
Peru	1	32	4

Note: TNPs = Total No. of Publications; TNCs = Total No. of Citations; TLS = Total Link Strength.

ranks countries based on their scientific contribution to the literature on the application of digital technologies and artificial intelligence in sustainable waste management. The ranking is based on three indicators: Total Number of Publications (TNPs), Total Number of Citations (TNCs), and Total Link Strength (TLS), the latter measuring the degree of connectivity and collaboration within the international bibliographic network.

China leads the ranking with 39 publications and a TLS of 23, confirming its position as one of the main global players in terms of both research volume and integration into international scientific networks. Egypt follows with 22 publications and a TLS of 18, highlighting its growing participation and strong citation impact (139 citations), a sign of increasing international recognition of its national scientific output. Germany, with a comparable number of publications (25), stands out for its highest citation impact (167 citations), indicating significant academic recognition of its contributions.

Countries such as Nigeria (TNP = 16) and Portugal (TNP = 14) also show active engagement in this research area, albeit with lower citation counts, suggesting potential for future growth. Medium-sized European countries like Spain (TNP = 9) and Hungary (TNP = 8) rank mid-table, while Saudi Arabia is notable for its high citation-to-publication ratio (90 citations from just three articles), reflecting a small but highly influential research output.

At the bottom of the ranking are India (TNP = 2) and Peru (TNP = 1), which, although still marginal in terms of quantitative contribution, show signs of entry into the global research network, with noteworthy TLS values indicating a degree of connectivity to the international scientific landscape.

To complete the analysis of bibliometric data on the geographic distribution of scientific output on this topic, a comparative table was constructed associating each country with three structural indicators. The selected indicators are as follows:

Per capita food waste (kg/year): estimated annual food waste per person, based on FAO (2021) [43] and UNEP (2021) [44] data;

Global Hunger Index (2023): a composite index that evaluates the severity of hunger at the national level [45];

Prevalence of undernourishment (%): the share of the population lacking access to adequate caloric intake, based on FAO data (SOFI Report, 2023) [45,46].

These indicators reflect both the inefficiencies of national food systems and levels of chronic food insecurity, enabling the bibliometric data to be interpreted within a broader socio-economic and environmental context [47].

Table 6. Structural indicators on food waste and food insecurity across selected countries. Authors’ elaboration based on data from FAO [43], UNEP [44], and GHI [45].

Country	Estimated Food Waste per Capita (kg/year)	Global Hunger Index (2023)	Prevalence of Undernourishment (%)
China	93	Low	2.5
Egypt	73	Moderate	5.5
Germany	75	Low	2.5
Nigeria	189	Serious	12.7
Portugal	93	Low	2.5
Spain	77	Low	2.5
Hungary	68	Low	2.5
Saudi Arabia	105	Low	2.5
India	50	Moderate	16.6
Peru	70	Moderate	7.0

highlights the pressures faced by national food systems. The data on per capita food waste reveals significant disparities: Nigeria and Saudi Arabia exceed 100 kg/year, indicating high levels of inefficiency, while India and Peru have much lower values, though within contexts marked by infrastructural fragility. The Global Hunger Index (GHI) reinforces this perspective: Nigeria is classified as being in a “serious” condition, while India, Egypt, and Peru fall into the “moderate” category. Even more striking are the figures on the prevalence of undernourishment, which reach 16.6% in India and 12.7% in Nigeria, signaling a concrete and persistent exposure to chronic food insecurity. These findings support the hypothesis of a non-linear relationship between critical food conditions and scientific engagement, while also revealing systemic asymmetries that affect countries differently in terms of both challenges and research capacity. These key insights are synthesized in Table 6.

3.3. Mapping Key Research Streams: Analysis of Top Cited Contributions

The bibliographic analysis then focused on the 20 most cited articles in the selected corpus. These contributions, presented in

Table 7. State-of-the-art analysis. Author’s elaboration.

Title	Authors	Citations	Main Focus	Research Trends
Evaluation of energy efficiency of various biogas production and utilization pathways	[48]	673	Energy efficiency in biogas production and use	Biogas pathways, energy balance, renewables
Greenhouse gas emissions from inorganic and organic fertilizer production and use: A review of emission factors and their variability	[49]	252	GHG emissions from fertilizer use	Circular fertilizers, composting, emissions
Big data for agri-food 4.0: Application to sustainability management	[50]	213	Big Data for sustainability in the agri-food supply chain	Industry 4.0, Big Data, sustainability
Exploring the application of Industry 4.0 technologies in the agricultural food supply chain: A systematic literature review	[51]	113	Industry 4.0 in agri-food supply chains	IoT, Big Data, cloud, food sustainability
Environmental balance of the UK biogas sector: An evaluation by consequential life cycle assessment	[52]	108	LCA of the UK biogas sector	Biogas, anaerobic digestion, LCA, UK
Using Artificial Intelligence to Tackle Food Waste and Enhance the Circular Economy: Maximising Resource Efficiency and Minimising Environmental Impact: A Review	[53]	101	AI applications to reduce food waste	AI, machine learning, CE, waste reduction
Critical success factor analysis of blockchain technology in agri-food supply chain management: A circular economy perspective	[54]	101	Success factors for blockchain adoption in agri-food	Blockchain, supply chain, CE, traceability
Global primary data on consumer food waste: Rate and characteristics—A review	[55]	100	Global consumer food waste data	Consumer behavior, food waste, SDG 12
Circular economy and solid waste management: challenges and opportunities in Brazil	[56]	95	Circular economy in waste management	Urban waste, CE, environmental impacts
Big data in the food supply chain: a literature review	[57]	89	Role of digitalization in the food supply chain	Blockchain, AI, RFID, supply chain resilience
A system dynamics model for evaluating food waste management in Hong Kong	[58]	87	Dynamic modeling for urban food waste	System dynamics, urban policies, food waste
The Future of Food	[59]	82	Future prospects for sustainable food systems	Bioeconomy, alternative proteins, digitalization
A decision support framework for the design and operation of sustainable urban farming systems	[60]	80	Decision support for sustainable urban farming	Urban agriculture, digitalization, efficiency
The “Vertigo” of the Food Sector within the Triangle of Climate Change, the Post-Pandemic World, and the Russian-Ukrainian War	[61]	77	Combined effects of crises on the global food sector	Multi-crisis, resilience, food geopolitics
Evolution and Current Challenges of Sustainable Consumption and Production	[62]	72	Sustainable consumption and production challenges	CE, degrowth, SCP indicators
Space, time, and sustainability: The status and future of life cycle assessment frameworks for novel biorefinery systems	[63]	65	LCA for novel biorefinery	Organic waste, bioenergy, environmental impact
Digital platforms: mapping the territory of new technologies to fight food waste	[64]	60	Digital technologies to combat food waste	Apps, collaborative economy, surplus management
Food waste recovery pathways: Challenges and opportunities for an emerging bio-based circular economy	[65]	55	Recovery strategies in food waste valorization	Zero waste, by-product utilization
A circular capability framework to address food waste and losses in the agri-food supply chain	[66]	53	Circular supply chain in agri-food	Digital twin, stakeholder mapping, CE
Life cycle assessment of microalgae-based aviation fuel	[67]	50	LCA for algae-based biofuels	Microalgae, biofuel, LCA

, represent foundational works that have laid the theoretical and methodological groundwork for the development of the field. They provide a comprehensive overview of the main research trends that have emerged in recent years, as well as insights into the most influential authors and journals.

The decision to focus on the most cited works, rather than the most recent, was aimed at avoiding a double temporal bias in the bibliometric analysis: on one hand, the time constraint imposed by the selected period in the Scopus search; on the other, the risk of favoring articles too recent to have made a measurable impact. This strategy, therefore, made it possible to concentrate on studies that have generated significant scientific impact, allowing the reconstruction of the main theoretical, operational, and technological pathways that are shaping the debate on digitalization and circularity in food systems [18,20,21].

The analysis of the 20 most cited articles reveals a body of literature focused on the efficient valorization of resources within agri-food systems, with growing attention to strategies that integrate digital technologies, bioeconomy approaches, and environmental assessment tools. Resources are explored not only as tangible materials, but also as data, information flows, and management processes to be optimized from a sustainability perspective.

A first thematic cluster concerns the evaluation of energy and environmental efficiency in the treatment of organic waste. The studies by Pöschl et al. [48] and Styles et al. [50] focus on the energy balance of the biogas sector, comparing different anaerobic digestion technologies and biomethane utilization pathways. In parallel, Walling and Vaneeckhaute [49] analyze the life cycle of fertilizers, highlighting the potential of nutrient recovery strategies to reduce emissions.

The concept of circularity as a resource management strategy is central to the contributions of Masi et al. [56] and Santagata et al. [65], who explore both the benefits and challenges of integrating food waste into circular economy models. Kusumowardani et al. [64], on the other hand, propose a digital framework for circular agri-food supply chains, based on tools such as digital twins and active stakeholder engagement.

From a bioeconomy perspective, studies such as those by Guo et al. [67] and Vance et al. [61] emphasize the role of innovative biological resources: from microalgae-based biofuels to the use of organic waste in advanced biorefineries, these approaches aim at the regeneration of matter and energy, fully aligned with circular economy principles.

The digitalization of the agri-food supply chain emerges as a strategic lever for improving traceability, resilience, and transparency. The works of Belaud et al. [50], Yontar [54], and Rejeb et al. [57] examine the adoption of big data, blockchain, and AI, while Li et al. [58] demonstrate the effectiveness of digital decision-support systems in urban agriculture contexts. These technologies act as enabling resources, essential for reducing inefficiencies and enhancing operational sustainability.

Finally, Galanakis [59,61] offers a broader reflection on the impact of global crises in reshaping food systems. The author stresses the need to build resilient systems, grounded in circular logic, new logistics configurations, and food waste reduction strategies as key levers for ensuring both food and environmental security.

The reviewed literature outlines a dynamic and rapidly evolving research landscape, where intelligent, regenerative, and integrated resource management is emerging as a strategic priority to tackle the environmental challenges associated with agri-food waste. The integration of emerging technologies, environmental assessment tools, and circular models provides concrete pathways for a sustainable transition aligned with the Sustainable Development Goals (SDGs).

3.4. Thematic Clusters: Synthesis and Interpretation

Overall, the reviewed literature reveals a growing focus on the integration of digital technologies, intelligent solutions, and systemic approaches to address the environmental challenges associated with agri-food waste. In particular, artificial intelligence (AI) emerges as a crucial enabling technology: it is used not only to predict waste flows and optimize logistics, but also to support environmental monitoring and the circular valorization of organic waste [50,53,60].

The evidence clearly points to an ongoing technological transition from linear disposal models toward predictive, integrated, and data-driven solutions that combine data analysis, operational efficiency, and environmental sustainability [48,58,60]. This trend aligns closely with SDG 9, which promotes the adoption of advanced technologies in industrial processes, and SDG 12, which targets the reduction in waste throughout the entire food supply chain.

Particular importance is also given to technologies such as anaerobic digestion, pyrolysis, blockchain for traceability, smart packaging, and the use of digital twins in management models, all of which contribute to the development of a low-emission digital circular economy, in line with the goals of SDG 13 [49,52,54,56,65,66].

This contribution adopts an integrated methodological approach, combining bibliometric analysis with a qualitative thematic review, applied to a corpus of 20 highly cited articles. This strategy enabled the identification of six key thematic areas, highlighting interdisciplinary connections among environmental, technological, social, and economic dimensions. Specifically, the analysis of keywords with co-occurrence ≥ 3 and Total Link Strength (TLS) ≥ 60 allowed for a consistent mapping of the main conceptual areas in the literature, offering an integrated and updated overview of the scientific landscape on the topic [55,57,63]. A summarized visualization is provided in

Table 8. Summary of emerging macro-themes and associated keywords. Author’s elaboration.

Macro-Theme	Representative Keywords	Authors
Predictive Technologies and Artificial Intelligence in Waste Management	machine learning, artificial intelligence, prediction, controlled study, optimization, algorithm	Pöschl [48], Li [60], Onyeaka [53], Toniolo [58]
Circular Economy and Sustainable Food Systems	circular economy, food waste, recycling, redistribution, sustainable development goal, review, article	Galanakis [60], Masi [56], Chen [65], Campos [66]
Digitalization and Traceability of Food Flows	digitalization, blockchain, digital twin, smart packaging, traceability, supply chain	Belaud [50], Yontar [54], Rejeb [57], Cane [64]
Valorization and Recovery of Organic Waste	anaerobic digestion, biomass, pyrolysis, enzymatic conversion, biorefinery, compost	Styles [52], Walling [49], Rorat [63], Batan [67]
Behavioral Factors and Domestic Context	consumer behavior, household, food loss, nonhuman, community influence	Dou [55]
Global Crises and Agri-Food System Transition	resilience, climate change, supply chain disruption, system transformation, food security	Galanakis [59,61], Glavič [62]

.

4. Discussion and Conclusions

4.1. Comparative Assessment of Key Digital Technologies for Circular Agri-Food Waste Management

Digital technologies play a central role in enabling the circular transition in agri-food systems. However, their levels of maturity, adoption, and effectiveness vary significantly across technological domains and application contexts [68]. While the bibliometric analysis revealed strong academic interest in artificial intelligence, blockchain, and digital twin technologies, these tools are not equally developed or accessible in real-world scenarios.

To provide a clearer and more operational perspective,

Table 9. Comparative overview of five key digital technologies applied to agri-food waste management, highlighting their main advantages, limitations, and indicative Technology Readiness Levels (TRLs). Authors’ elaboration based on data from Yfanti and Sakkas [70].

Technology	Main Advantages	Limitations and Challenges	Indicative TRL
AI	– Waste flow forecasting – Logistics optimization – Decision-making support	– Limited algorithm transparency – Requires high-quality historical data – Adoption complexity	7–9
Blockchain	– Secure and transparent traceability – Data integrity – Automation via smart contracts	– High implementation costs – Limited scalability – High energy consumption (in some protocols)	5–7
DT	– Real-time simulation and forecasting – Virtual modeling of processes and resources	– Complex data integration – High infrastructure costs – Requires advanced expertise	4–6
IoT	– Real-time monitoring – Process automation – Environmental data collection (e.g., humidity, temperature, weight)	– Cybersecurity risks – Interoperability challenges between devices – Maintenance costs	8–9
Big Data Analytics	– Large-scale data analysis – Resource optimization – Strategic decision support	– Requires robust infrastructure – Data quality and variety may vary – Interpretation complexity	7–9

offers a comparative overview of five key technologies—artificial intelligence (AI), blockchain, digital twins, Internet of Things (IoT), and big data analytics—frequently cited in the selected literature. The table summarizes their main advantages, current limitations, and indicative Technology Readiness Level (TRL) in the context of agri-food waste management [69].

This comparative framework serves as a starting point for the discussion that follows, which critically explores implementation challenges, stakeholder resistance, ethical and governance dimensions, and the need for contextualized field validation.

4.2. Empirical Validation: The Winnow Case

To reinforce the practical relevance of the findings emerging from the bibliometric analysis, this section presents an illustrative and internationally recognized empirical validation: the case of Winnow. Winnow is a technological platform that leverages artificial intelligence, sensors, and computer vision systems to monitor and reduce food waste in professional kitchens such as those in restaurants, canteens, and hotels [71].

By automatically identifying discarded food and algorithmically processing the data, Winnow enables real-time traceability of food waste types, facilitating more efficient purchasing decisions and reducing surplus. The results achieved across more than 3000 kitchens in 94 countries are significant: an average food waste reduction of up to 50%, food cost savings ranging from 3% to 8%, and a return on investment between 200% and 1000% within the first year of implementation [72]. In a pilot project conducted across 13 hotels in the United Arab Emirates, the system led to a 76% reduction in pre-consumer waste and a 55% reduction in post-consumer waste [73]. Globally, the technology has helped recover more than 60 million meals, preventing over 106,000 tons of CO₂-equivalent emissions annually [73].

The Winnow case exemplifies how the principles identified in the bibliometric clusters translate into operational practices in agri-food waste management. It thus serves as an indirect empirical validation of the theoretical trajectory described in this study, confirming the feasibility and scalability of digital solutions in supporting the achievement of SDG 9, SDG 12, and SDG 13.

4.3. Implementation Challenges and Stakeholder Resistance

While the analysis highlights the enabling role of digital technologies and circular economy models, it is essential to acknowledge the recurring barriers in real-world applications—particularly among small and medium-sized enterprises (SMEs), farmers, and local actors within the agri-food value chain. In addition to the general obstacles previously discussed, several studies point to forms of resistance rooted in cultural, organizational, or behavioral factors. These include low levels of digital literacy, perceptions of complexity associated with new tools, and skepticism toward data- and algorithm-driven decision-making systems [74].

For example, small-scale farmers often express doubt about the use of AI-based predictive systems, which they perceive as poorly adapted to local climatic or economic conditions. Similarly, many SMEs are hesitant to adopt blockchain-based traceability technologies due to the required investment and uncertainty regarding tangible economic benefits [75]. In some cases, the lack of co-design processes in digital systems has resulted in solutions that are misaligned with the actual operational needs of end users. These findings suggest that for technological innovation to be effective and scalable, it must be accompanied by participatory models, inclusive governance, and targeted training programs aimed at reducing access and competence asymmetries.

4.4. Ethics of Digitalization and Governance Issues in Less Developed Contexts

Beyond technological and environmental considerations, the digital transition in agri-food waste management raises critical ethical and governance questions, particularly in less developed regions or in contexts marked by deep structural inequalities. The adoption of digital technologies often entails the collection and processing of large volumes of data, frequently without adequate guarantees of transparency, user control, or respect for data sovereignty [75].

In many Global South contexts, data ownership tends to remain concentrated in the hands of external technology providers or global actors, thereby reinforcing patterns of dependency and asymmetry [76]. Compounding this issue are significant barriers to equitable access to digitalization, driven by economic, infrastructural, and cultural constraints. In the absence of targeted capacity-building initiatives and inclusive regulatory frameworks, there is a real risk that digitalization may create new forms of exclusion or exacerbate existing divides—undermining the sustainability and social justice goals promoted by the 2030 Agenda [76]. It is, therefore, imperative to embed principles of data ethics, local stakeholder participation, and distributive equity into the design and implementation of digital solutions.

4.5. Strategic Alignment with the SDGs and Systemic Equity Issues

An original contribution of this study lies in bridging a well-structured bibliometric overview with a critical interpretation of sustainability goals and systemic disparities. While previous research has explored the role of digital tools in the agri-food sector, few studies have provided an integrated analysis that maps enabling technologies to SDGs, while also highlighting global knowledge asymmetries. This work helps fill that gap by offering a dual perspective, technological and geopolitical, on digital circularity.

The analysis carried out highlights how the integration of digital technologies and circular economy strategies represents a key enabler in the transition toward more sustainable and resilient agri-food systems. Technologies such as artificial intelligence (AI) [51], digital twin systems [64], blockchain [52], and big data platforms [48] are making significant contributions in preventing, monitoring, and valorizing food waste, promoting production models aligned with SDG 9, SDG 12, and SDG 13.

To provide a clearer and more operational view,

Table 10. Contribution of key digital technologies to selected Sustainable Development Goals. Authors’ elaboration.

Digital Technology	SDG 9	SDG 12	SDG 13
AI	Promotes innovation in monitoring and forecasting processes	Reduces food waste and optimizes material flows	Supports data-driven decisions to lower emissions
Blockchain	Strengthens distributed digital infrastructure	Enhances transparency and traceability across the supply chain	Facilitates sustainable resource management
DT	Enables simulations and virtual system optimization	Assists in the efficient management of resources and waste	Models the environmental impact in predictive scenarios
IoT	Builds real-time monitoring infrastructure	Makes the life cycle of products and waste traceable	Enables environmental monitoring and alert systems
Big Data Analytics	Supports evidence-based innovation	Optimizes resource use through predictive analytics	Helps identify climate-related patterns and scenarios

presents a visual matrix mapping the five key digital technologies identified in this study to the SDGs they most directly contribute to. This mapping is based on recurring themes and applications found across the selected corpus and highlights the complementary contributions of each technology to sustainability targets.

The bibliometric analysis revealed two main thematic streams in recent literature. The first focuses on the adoption of digital technologies for waste management, including waste flow prediction [58], logistical optimization [55], product traceability [52], and real-time environmental monitoring [50]. The second stream centers on the valorization of organic waste through technologies such as anaerobic digestion [47], pyrolysis [61], and the use of residual biomass [65], emphasizing their potential to close material loops, recover valuable resources, and reduce greenhouse gas emissions.

Although often grouped under the broad umbrella of waste-to-energy strategies, anaerobic digestion and pyrolysis differ significantly in terms of technological suitability and contextual applicability. Anaerobic digestion is particularly well-suited for wet, easily degradable organic matter such as food residues, animal manure, and crop processing slurries. It has gained traction in several European countries thanks to favorable policies and biogas-related incentives. In contrast, pyrolysis is more appropriate for dry, lignocellulosic biomass, such as straw, prunings, nutshells, and woody materials, and requires greater thermal control, investment capacity, and a market for by-products like biochar or syngas. In low- and middle-income settings, anaerobic digestion tends to be more accessible and scalable, while pyrolysis may support more advanced circular bioeconomy strategies where industrial capacity is available.

An original contribution of this study lies in the integration of bibliometric analysis with a critical interpretation of structural indicators, which revealed mismatches between scientific output and food system vulnerabilities. Some African and Middle Eastern countries, such as Nigeria and Egypt, stand out for their high academic activity despite facing significant food insecurity, suggesting a reactive orientation of national research agendas [54,63]. Conversely, countries like India, although affected by high levels of malnutrition, remain underrepresented in international scientific production, likely due to systemic, infrastructural, or linguistic barriers [53]. These findings point to the need for promoting epistemic equity in global knowledge production on food waste, through policies that support contextualized research in low- and middle-income countries. In this regard, digital technologies can serve not only as operational tools, but also as drivers of democratization in resource management and a reduction in systemic inequalities [59].

There is also increasing interest in circular business models, enabled by digitalization, that promote traceability, transparency, and resilience throughout the entire supply chain. The adoption of predictive algorithms [51] combined with smart sensors and IoT systems [55] allows for dynamic inventory optimization, surplus prevention, and more efficient food redistribution [62]. These approaches directly contribute to achieving SDG 12 by fostering more efficient use of natural resources and promoting sustainable production and consumption models. The implementation of interconnected and intelligent digital infrastructures also reflects the principles of SDG 9, which aims to advance sustainable industrial innovation [49]. Furthermore, the reduction in emissions from organic waste disposal and the generation of renewable energy through biological systems reinforce the relevance of these solutions for SDG 13 [47,61]. Despite these promising developments, these contributions are not yet equally reflected in global practice. Strengthening the operational link between technology and policy, particularly in underrepresented regions, remains a strategic priority.

5. Study Limitations

Despite the integrated approach adopted and the robustness of the sources analyzed, this study presents some methodological limitations that should be acknowledged.

First, the bibliometric analysis was conducted exclusively using the Scopus database, one of the most authoritative platforms in the scientific domain. However, the exclusion of other relevant sources such as Web of Science or IEEE Xplore may have limited the representativeness of the dataset, particularly with respect to emerging research on digital technologies or interdisciplinary contexts.

Second, the priority given to technical-scientific contributions may have led to the underrepresentation of literature from social, political, and economic sciences, which could offer complementary perspectives on governance, environmental justice, and the real-world implementation of intelligent technologies in waste management.

Another limitation concerns the static nature of the analysis, which focused on highly cited articles within a defined time frame. This approach risks overlooking recent innovations that are rapidly evolving but not yet consolidated in high-impact literature.

The structure of the study is primarily conceptual and descriptive, based on documentary evidence and semantic mapping, without direct empirical validation. Specifically, the analysis lacks applied studies or field data that quantitatively assess the real-world effectiveness of AI and digital twin technologies in agri-food or urban contexts.

To address these critical aspects and strengthen the theoretical and operational maturity of the field, several priority research directions are proposed, as summarized in

Table 11. Future research directions in intelligent food waste management. Own elaboration.

Research Area	Key Issue	Guiding Research Question
Empirical validation	Lack of real-world testing of AI and DT in food waste management	How do AI-based systems perform in operational settings in terms of cost-efficiency, scalability, and impact?
AI and policy integration	Limited understanding of how digital tools support governance and regulation	In what ways can AI and digital twins enhance circular economy decision-making in public policy?
Algorithmic performance comparison	Unclear effectiveness of different AI models across waste types	Which predictive models are most suitable for specific categories of agri-food waste?
Socio-ethical implications	Insufficient exploration of social acceptance, equity, and workforce impact	What ethical, social, and institutional challenges arise in the adoption of AI and DT in the waste sector?
Bibliometric expansion	Exclusion of emerging technologies like edge AI or generative AI	How can bibliometric mapping be extended to capture fast-evolving innovations in digital circular waste systems?

.

This research roadmap represents a first step toward methodological and practical consolidation of digitalization in agri-food waste management systems. Addressing these questions will be essential to fostering the development of sustainable, scalable, and inclusive models, aligned with the priorities of the 2030 Agenda and the ongoing technological evolution.