 Thank you for all those who are attending today's seminar. Please, there is session of the Alkmetarismic seminar series here in person or online. Professor Perotti asked me to extend her greetings and her regrets to you all. Since she wouldn't be able to be here due to child dreams. Sesson neskratenim... V sessonu neskratenim je oto ozvrte, ko je se z Prof. Perotti. Vespoje, če sem se neskratenih iz vsev, in se odmah vzelo v erbano-sociali koronjalistice pomečenih oben, in začasnjeti smokovom režim tudi, če potem sem preforjeno komputationalno analizične in drugiratične režime v smetniej delovosti z medustvenosti. Zato v tem, ki sem prflike, zima vedem, da sem počkama tudi kot ampak societi. Jel sem postočeno o predstavitje na projekti? vzvečenih je Alejandra Acevedo, kaj je vzvečen na... Futsikerti. Futsikerti? Na narodilizmu. Elizabeto, kaj je vzvečen na projekti, kaj je vzvečen na vzvečenih projekti, kaj je Zanis vzvečen vzvečen na projekti. Promet, kaj je vzvečen na sistemikologiji, vzvečenih, vzvečenih, vzvečenih. Promet, kaj je Voskoli? Voskoli, vse zelo, na pravda, vzvečenih. Spokoj, vzvečenih vzvečenih na vzvečenih in vzvečenih. Vzvečenih, vzvečenih, erbarno celo naseljevac na smetnih in izvahovah za urbarnostične učeščenje. Usovstvo in stavljevac na ječkih z konceptu urbarnostičn weniger imeli pendento pristки potenči, povoljnoch prikodnanje, zatajnjih pravbi in prikodnjih, Taj potencijal je vseč izgleda, z školorami, vseči v tem vseči vrbani političnih vseči. In vseči, vseči političnih vseči vseči vseči vseči vseči vseči vseči vseči, as a framing device for reconcovering of the political character and underlying power dynamics, specific to the resource exchanges to which cities interact with nature. Otsit, the field of urban political ecology, the urban metabolism has been je bilo vsevstavljena z sistemov, tudi način, da je koncept, ki je zelo se početno zelo, način je vsevstavljena z urmanj, socijali, kologičnih sistemov. Zelo izgledaj se, da je bilo vsevstavljena z urmanj, zato vsevstavljena z urmanj, je bilo vsevstavljena vsevstavljena v tem framework of industrial ecology. Zato, koncepts, framework and theories about social-ecological systems that are not directly linked to the urban metabolism notion but that are relevant to research on urban metabolism have been developed in fields that are also systems based. Urban ecology and social ecology. Notable exception, notwithstanding, system-based approaches have generally devoted limited attention to the agentic dimensions of the urban metabolism as a concept. I.e. they have devoted limited attention to identify the different forms of agency, operating in urban social-ecological systems. In contrast to systems-based approaches to the urban metabolism, urban political ecology inspired accounts of metabolic exchanges have questions, the relations between different forms of agency that operate in cities and are part, are constitutive of urban metabolic exchanges. In particular urban political ecology has engaged in critical dialogue with conceptualization of agency coming from philosophy and social sciences. In particular they have engaged with new vitalist ontologies and the actor-network theory against which they have generally argued that advice should be maintained between human and non-human forms of agency. Despite the diversity that characterizes urban political ecology inspired accounts of the urban metabolism, those such accounts share a minimum common denominator which reduces to three core critical questions that can be asked to clarify what is meant by agency in relation to the urban metabolism. The first question is whose agency and more specifically what systems, processes, entities, etc. can be identified as the actor possessing the agentic capacity in question understood as the capacity to exerting certain influence and effect a certain form, a certain kind of change. The second question, critical question is what is acted upon, i. what systems, processes, entities, etc. constitute the target of the actor's influence. Third, how and when do the actors in question exert their agentic capacities, i. what are the concrete special-temporary and special-temporary situated events, processes, etc. by which actors exert their agentic influence and effect change on other actors or on their surroundings. Those critical questions constitute the backbone for establishing a conceptual, for outlining a conceptual framework that characterizes agentic features of urban sociological structures and processes by connecting them with agentic dimensions constitutive of the urban metabolism concept itself. More specifically, this framework, this conceptual framework this conceptual framework characterizes agentic features of urban sociological structures and processes by showing how those same features embody kinds of agencies that determine how the urban metabolism concept is defined and understood. Therefore, forms of agency that are constitutive of the meaning of the term urban metabolism and of the structure of the urban metabolism concept. Starting from those questions, our article outlines agentic dimensions at the core of the urban metabolism concept and shows how such dimensions are involving structures and processes that comprise the complex urban sociological systems. Particularly, we analyze a corpus of journal articles dealing with the urban metabolism and with the theoretical foundations of systems-based approaches to the urban metabolism or relevant to the urban metabolism i industrial ecology, urban ecology and social ecology. From this literature we extract concepts that describe different forms of agency and are cognate to the urban metabolism notion insofar as they are defined as the principles belonging to the semantic field of urban metabolism. By semantic field we refer to a set of terms with a closer relation in meaning of which can be subsumed under the same job label in our case the urban metabolism term itself. Our approach to extracting concepts from the literature relies on geo assumptions. First assumption, the idea that concepts can be understood as the meanings expressed by corresponding terms and the second assumption is a distributional view of the meanings expressed by those same terms. I view according to which semantic properties of words can be at least partially determined by analyzing the distributional properties such as the concurrence patterns between those same words and other words in particular in selected textual units. Based on those assumptions and drawing on text-mine in computational linguistic methods we identify the urban metabolism term semantic field as conveyed by our corpus and map it via semantic network. More precisely via four distinct semantic networks as we will see in the reminder of this presentation. Finally we use the conceptual framework just offline to highlight how the cognitive concepts is expressed by the relations between terms mapped by those semantic networks define fundamental agentic dimensions that are constitutive of the urban metabolism notion from a system-based and sustainability-oriented perspective. Moreover, we spotlight how identifying such identic dimensions contributes to advancing the urban metabolism concept as a tool for promoting urban sustainability more specifically sustainability of urban social ecological systems by illuminating the agentic features specific to elements that are constitutive of those same subsystems in particular specific to subsystems and processes to be identified within urban social ecological systems under certain conditions. Now I'm going to delve into the methodology results discussion that constitute the main body of the manuscript before drawing some conclusions. As for the methodology we started by collecting a body of 24 sources i.e. literature reviews or conceptual papers explicitly mentioned in the term urban metabolism and also additional texts related to either one of the systems approaches to urban metabolism to urban social ecological systems discussed that I mentioned or a mixture of those same approaches. Here is the database of the 24 sources identified. For instance, we have this document which is part of a book that surveys the conceptual and methodological foundations of the social ecology approach. We also included literature reviews about dealing with urban metabolism. For instance, this literature review that deals with the use of urban metabolism as a research framework for studying sustainability and this was just an example of the kind of documents that we used to constitute the corpus that we analyzed and based on those 24 sources we identified 1481 relevant journal articles that were selected by browsing the bibliography, the references section of the 24 documents which were screened in order to identify only select research articles, editorials or literature reviews in English relevant to the content of the corresponding source document rather than only to the methodology used in the source document. By doing so, we built corpus of documents published between 1899 and 2023 with 82% of them issued in the 2003-2023 period. And we also categorized those documents according to the scientific field by looking at relevant features such as, for instance, title, first author, publication venue, author or keyword content and the source document mentioned in them. And as you can see, there is quantitative bias in industrial ecology which constitutes the main disciplinary affiliation of the document we collected and which is also and this bias is related to the fact that industrial ecology is the field in which the term or metabolism and the concept of urban metabolism are more widely used in opposition to the other systems based approaches in which social ecological systems in which we were interested. Based on this selection, we extracted from scopus and Google scholar the abstract titles and when available author or index keywords of the references that are selected. And we assembled those elements into documents pre-processed through our packages Udepipe and Contida to abide by standard pre-processing practices. Therefore, all such documents were homogenized, grammatically annotated using a part-of-speech tagging algorithms tokenized to extract individual award tokens contained in those documents lematized etc. And subsequently we used the R package Contida to join possible multi-word expressions in the form of n-grams phrases of n-conceptive tokens occurring in the same document and identified as lighting window filter based on the tokens we previously selected and reviewed. In this way each document was represented as a bag of 1 to 5 n-grams a collection of monograms in the other tokens big-grams, two-word expressions etc. up to pentagrams, five-word expressions and this is for instance so this you can see on the right side the screen an example of of a document so that we constituted by pasting title, abstract and available keyword document was represented as a bag of 1 to 5 n-grams that you can which you can see an excerpt on the other side so to the left side represent the left side document represent actual input that we used to perform our analysis and we performed a statistical test on the n-grams we had we had built to minimized the risk of including meaningless expressions in our selection and we performed some additional cleaning by removing really frequent n-grams and meaningless or redundant n-grams we end up with lexicon of 13 unique terms and we analyzed by applied the methodology that is in this figure so we started by we split our corpus into two subcorpora a subcorpus of reference and a subcorpus of interest subcorpus of reference included sorry, subcorpus of interest included only the 251 documents containing the term of metabolism itself whereas the subcorpus of reference contained only contained the remaining 1230 documents and by by split the corpus in this way we were able to adopt computational corpus driven linguistic approach to the analysis of the differences between those two subcorpus that rely on a particular specific log likelihood test statistic to assess how statistically meaningful the differences between how statistically significant the differences between frequencies of words in the two subcorpus are more specifically we used the subcorpus of reference as general backdrop to test the specificity of the words contained in the subcorpus of interest specificity of words occurring with the term of metabolism by testing the significance the statistical significance of the differences between the relative frequencies of those corpus words in the subcorpus of interest and the subcorpus of reference those statistical differences can be regarded as quantifying the linguistic significance for kindness of the term considered for characterizing the content of the documents comprised by the subcorpus of interest as opposed to their statistic features statistic features of those same documents the analysis of such differences that provides basic insight into how the terms in questions constitute a semantic field as previously defined is clear other questions about this first logical aspect by doing so we identified 271 significant term keywords and key multiword expressions in the form of engrams and based on this selection we identified positive keywords and key multiword expressions keywords and key expressions whose frequency in the subcorpus of interest was statistically higher than their frequency in the other subcorpus in a statistically significant fashion and negative keywords and key multiword expressions ki words and key multiword expressions are underused in the subcorpus of interest so the positive significant terms can be the positive significant terms thus identified can be regarded as constituting the core of the semantic field associated with the term urban metabolism whereas the other subcorpus negative words can be understood can be regarded as tracing the boundaries of this same semantic field i.e. identifying fault lines in which the semantic field of the urban metabolism of the term urban metabolism can be seen as intersecting with the semantic fields of other terms present in the corpus we analyzed once we extracted our 271 significant terms we adopted what we computed various indicators to assess the relations between those same terms for instance we computed the Simpson coefficient who is a measure of the strength of the coherence between two terms from a theoretical standpoint we also computed the term frequency of those terms and our significant terms and this other indicator allows to measure the relevance of a certain term T with respect to a document in which the same term is contained third we computed and arrived the disciplinary score for each significant term document pair based on the corresponding term frequency inverse document frequency to summarize roughly we used the term frequency for each term document pair we used the corresponding tfidf value and disciplinary score attributed to the document in question based on its disciplinary affiliation for instance industrial ecology score of one if the document was affiliated if the documents only disciplinary affiliation was to industrial ecology one half if it was if the document was affiliated to industrial ecology and another discipline etc and by weighing the corresponding tfidf values by the disciplinary affiliation score disciplinary affiliation scores attributed to the we computed the disciplinary scores for different for each score for each term and the affinity of this term for a particular discipline and fully the relative frequency or more specifically the relative yearly frequency of a term which measures this term frequency relative to the yearly question i.e. the share of the ratio between the frequency of the term in question for a given year and with respect to the total number of terms contained in the documents published this same year and using the ER package WGCNA we computed four payways spearman rec correlation coefficients one for each indicator four pairs of significant terms the set of coefficients that is calculated for each pair of terms measured the ordinal similarity of those same terms of course patterns along the four dimensions in particular of course with other terms in the case of pairs correlation between simpson coefficients of course irrelevant documents in the case of pairs correlation between tfidf values of course in classes of irrelevant documents assigned to the same discipline or disciplines in the case of correlation between aggregate disciplines course and of course across time in the case of correlation between relative yearly frequencies finally we average over all those sets of four payways correlation coefficients for each pairs of significant terms in order to synthesize this wealth of information and this way we computed an average payways correlation matrix for our significant terms in particular notably intermediary of fissures zeta transformation which is mathematical transformation that is widely adopted for comparing and evaluating correlation coefficients by mapping correlation coefficients over almost normally distributed variables and and back and and then mapping these the valid instances of this of this quasi normal distribution back to the original equation so we adopted this methodology for three reasons first because we wanted to extract terms in the form of keywords or key multi-word expressions that significantly characterize a specific section of our corpus second we want to identify in such a set of significant terms groups of words and multi-word expressions that appear to be similar along multiple dimensions and third we wanted to amplify the similarity among the elements of those in groups by averaging correlation coefficients and once we computed once we analyzed we computed all those average correlation coefficients we we operated the further selection on the second terms we identified in particular we split those terms in four categories based on three points that corresponded to values of the pair words between those terms and the urban metabolism term specifically you here you can see on the screen the distribution of the relevant terms with respect to their pair words coefficient with urban metabolism and terms that whose coefficient value is lower than 28 constitute the lower terms that are constitute negative quote-unquote negative significant terms i.e. terms that are that are only weakly associated with urban metabolism using the Simpson coefficient as a measure of association terms whose Simpson coefficient values are are between 28 and 50 constitute the lower tail of group of terms that are mildly associated with the urban with term of metabolism terms whose Simpson coefficient values are fall within within a range going from 50 to 80 constitute the upper tail of terms that are still mostly mildly associated with urban metabolism whereas terms whose piegrist Simpson coefficient with respect to the urban metabolism term is greater than 80 constitute term that are closely the term urban metabolism and therefore can be regarded as constitute the core of the semantic field associated with the urban metabolism term by operating this selection we identified four groups of terms that we used to to plot four semantic networks each corresponding to a layer of the semantic field associated with the term urban metabolism in particular for for each one of those four group we represented the terms contained in that group as nodes in the corresponding semantic network we only retained positive and strong positive, strong statistically significant pay-wise correlation coefficients on terms more specifically correlation coefficients that were greater than 0.95 then we used those correlation coefficients to establish a list of edges for each network then we performed a clustering analysis on the different networks using the urban community detection model and for each cluster in each network we then selected up to eight influential nodes based on their eigenvektor centrality which is a measure of influence that considers both the degree by the number of connections of a given node and the degrees of its neighboring nodes in this we we managed to plot those four networks in this in this in this this is the first network which represents the terms that are weakly associated with urban metabolism and constitute the four boundaries of the of the of the semantic field in this term as you can see this network consists of four clusters and is structured by the term biodiversity human, biophysical science and biodiversity in this in this in this in this biodiversity, human, earth science and biophysical so to go to interpret this we identified documents that are likely to provide relevant context to the meaning of the different terms either at the network level or at the cluster level more specifically for each for each bearer sorry for each cluster for each network and for each cluster in each network we score the relevance of all the documents contained in the corpus with respect to the networks or the cluster structure in terms in the following way we selected each structure in terms term frequency in direct document frequency values we weighed those values by the total number of relevant structure in terms contained in the corresponding documents and then we computed an average value for each document and ranked the document in the sending order in this way we built a sort of interpretation matrix for each network and each class and network we recorded the corresponding most influential terms and we associated relevant documents calculated according to the scores computed as I just mentioned so documents either from the corpus or document from the sub corpus of interest and document from the sub corpus of reference so for instance 2 to answer your question the term Baltimore is refers to the city and is the reason why it features in this cluster because this cluster is associated with series of studies that were conducted in the city of Baltimore in the urban area of Baltimore because for instance relationships between urban and greater Baltimore in region exactly more like Baltimore, the most full of talk because Baltimore has been the site of a long term ecological research project conducted within the framework of a particular school within the field of urban ecology that is relevant for our analysis of the urban metabolism concept because this school provided and developed a framework for studying urban social ecological systems most specifically for integrating the social aspects and the biophysical ecological aspects of ecosystems that is partially comparable partially alternative to how the to how urban social systems have been conceptualized for instance within the field of industrial ecology. This network comprises four clusters and the dominant clusters are cluster four and cluster one cluster four because it contains biodiverse biodiversity which is the most influential term represented by the most influential nodes in the network cluster one because it contains the remaining four the remaining four structural term and if we go back to the interpretation matrix documents that are listed among the 10 most relevant documents from the whole corpus to the structural term that characterize this particular cluster and which contains the term biodiversity are for instance Martinets Allier 2009 social metabolism and ecological distribution conflicts and languages of violation. Martinets Allier et al. 2010 social metabolism and ecological distribution conflicts and valuational languages and Greeks et al. Sustainable development goals for people and planet and Chandel and et al. 2018 global material flows and resource productivity 40 years of evidence If we look I was wondering for example there is this big one from Harvard that is basically an environmental standard input and cost analysis of emissions and I don't know because it's basically see the trade of several countries and see the environmental embedding part of it which is started like a series of but I don't think it considers diversity in this particular paper Yeah, because I was wondering what were the links of these paper stores is work by diversity I can show you more in that That's more the semantics of the world So the world just to maybe clarify a little So the documents here contain at least one of those words and the ranking of the words is based on the influence the degree of influence within the network itself but the degree of influence is calculated based on the correlation coefficients which considers not only concurrence in the same document but also similarity of patterns for instance if two words are used never concur but they are always sorry but the two words never concur but they are always used in documents of the same disciplinary affiliation there will be a connection that will show up between those two words in indirect connection mediated through the disciplinary disciplinary affiliation but in those the documents they all contain at least one of those words More precisely so for instance the third biodiversity and the the content of each cell shows the documents in which the row column term and those documents are listed in order in a ranking order based on the average relevance score for the two terms for instance biodiversity of all the documents or the most relevant documents that contain biodiversity are those listed in this cell and they are the ones that were mentioned before and Hartwig-Peters they show up with respect to environmental policy and nation and basically they connect and Hartwig-Peters is looking to connect environmental policy and nation that provide context on the relation between those two words in the framework of the cluster and for instance to go back to the term biodiversity so the four most relevant terms to understanding the meaning of term biodiversity within the context set by cluster 4 are the one that was mentioned before Shandon the two Martinets-Allier and Greeks the two Martinets-Allier 2009 and 2010 they connect biodiversity with extraction nation and world specifically they use concept from social ecology such as the concept of social metabolism and human appropriation of net primary production to frame biodiversity as the object of distribution conflicts technological goods and services that poor nations nations from the global south must face in a globalized economy whereas Greeks link biodiversity environmental policy and nation in discussing the preservation of Earth's life support as a priority for the United Nations sustainable developmental goals and more specifically Greeks use the concept of planetary boundaries parameter values describing the limits beyond which anthropogenic perturbations of core planetary processes threaten the planet Earth's global ecological stability and through this concept they characterize the rate of biodiversity loss as a core Earth system process that is subject to anthropogenic influence and finally Shander Tal I like the region between biodiversity and the remaining four and the four remaining TASR4 structure in terms in the analysis of a dataset tracing global material flows from 1970 to 2010 as support for environmental policymaking by tracing through this analysis of global material flow the authors account for the rate of biodiversity loss in countries from the global south especially Asia through the concept of material extraction i by attributing by ascribing this loss to a worldwide acceleration in the extraction trade and use of materials such as for instance fossil fuels or metal bores so this is an example of how a cluster can be interpreted based on the structure in terms the relation between structure in terms and the context providing documents for each network to be interpreted the two dominant cluster and if necessary other clusters and then we use the three core questions outlined at the beginning of this presentation to identify how the interpretation of the clusters allows characterizing the forms of agency that can be seen as being captured and described by the network but by each one of the different form networks and I'm just to skip to the end I'm going to show you the four rapidly in the control so this is still a network 4 sorry, network 1 cluster 1 sorry, cluster 4 is the one I just described rapidly cluster one is the other dominant cluster and basically cluster cluster 1 and cluster 2 can be seen can be interpreted as identifying and integrating integrated earth system and complex ecosystem at different scale as the target of human-led planetary social-ecological processes for instance international trade there are documents associated with those two clusters for instance I just mentioned explicitly tie such processes to changes at the global planetary level that are generally associated with the concept of the anthropocene therefore this connection the connection between the concept of anthropocene and the kind of processes affecting the integrated earth system that are described by the documents associated with the two cluster suggest a view of humanity as an ecological force that contributes to shaping an integrated earth system and we can thus infer that network 1 captures the anthropocene dimension of human agency i.e. the form of agency exerted by humanity albeit not as a homogeneous whole as an ecological force operating within and constitutive of the integrated earth system and then cluster 3 is very marginal but cluster 2 provides relevant information for instance regarding the particular mechanisms and events and processes through which this anthropocene agency is exerted in particular the term disturbance through its association with particular context providing documents allows to draws a connection between the concept of ecological disturbance as it has been developed within the framework of urban ecology and the way in which the term disturbance is used in some in studies about the urban metabolism to frame the alterations of metabolic processes within urban ecological systems due to imbalances in the through flow of materials and energy analogous to metabolic disorders therefore this concept of disturbance in two very different contexts when context more inspired by the framework conceptual framework of urban ecology and devoted to the analysis of ecosystems at different ecosystems as social ecological systems but at different scales and different circumstances let's say and on the other hand the term of disturbance is used within the concept of industrial ecology specifically with respect to urban metabolism and the connection between those two the way in which the term disturbance is used in those two different contexts allow to draw up on suggest that this term can be seen as referring to functional and structural alterations in processes that are specific to cities as heterogeneous yet integrated natural social systems and therefore this term disturbance within the context of cluster 2 can be used to characterize the way in which this anthropogenic anthropogenic agency is exerted at different levels from the level of cities to the level of the whole earth and more specifically by virtue of the fact that cities as urban social ecological systems are embedded in broader social ecological systems and by virtue of the fact that the metabolisms of urban social ecological systems can be seen as being embedded in the metabolisms of wider systems for example regional systems national systems and even more general processes that can be either social or socio-economic for instance the global flows of material energy driven by international trade or very broad and general bio-geochemical cycles Am I going to just show what we do in the other 3 so network 2 is represents the relations between turns that constitute the lower tail of the turns mic associated with the old metabolism consist of 4 clusters it's structured by term waste urban system material flow sustainable urban sustainability and the dominant clusters are cluster 3 here left cluster 2 on the right network 3 which corresponds to the upper end of the group of documents of turns mindless associated with urban metabolism comprises 3 clusters it's structured by the turns city, flow, urban resource metabolic and flow city and dominant cluster is cluster 1 which contains 3 out of 5 structure in turns network 4 which constitute the core of the which represents the core of the semantic field associated with terminal metabolism comprises 5 clusters and it's structured by the term metabolism urban metabolic flow metabolic metabolism research metabolism assessment the dominant cluster is the central one which consists of the nodes represented by urban metabolism urban metabolism research and metabolism assessment do just give you a sense of how the interpretation of those 4 clusters can you does not come characterization of the different subjective dimensions in the literature so this interpretation matrix might need some update but so for instance this first row is basically summary of demonstration that I of the interpretation that I presented that I just presented I'd say already and so the network network 1 characterize the anthropocene agency of humanity and the concept of cross-scale ecological disturbance is expressed by the cluster cluster 4 cluster 2 it will mismatch because I the list of words should be updated but it's the cluster 2 I showed before so this concept of cross-scale ecological disturbance is central in characterizing this particular aspect of humanity's agency of humanity's anthropocene agency network 2 characterizes different kind of agencies underpinning urban metabolism based assessments of the sustainability of social and ecological systems in particular the conceptual agency that is exerted by factors that can be identified and quantified by by indicators for instance indicators that rely on the conceptual framework of material flow analysis and the other kind of the form of agency that is captured by this network is the concrete agency that is exerted by the policy makers practitioners stakeholders that contribute to selecting those indicators and applying them to promote the sustainability of cities and social and ecological systems in general and we are going to skip to the last network 4 which through the concept of urban metabolic system allows to define particular form of agency that is relevant to understanding how the concept of urban metabolism is defined and used in literature will be analyzed and it is the cybernetic agency that urban social and ecological systems components exert as metabolic actors on one another and directly to all social and ecological systems in which they are comprised since this structure constitutes to the extent to which these structures constitute an ecological hierarchy so this aspect of the agency the metabolism of urban social and ecological systems is tied to a series of research in industrial ecology which uses energy based models models based on the concept of embodied energy as a universal equivalent that can be used to quantify the exchanges of energy in different forms and which is tied to a particular approach to the study of ecological systems in general and this is the approach developed by Howard Odum that I mentioned in last month's presentation and therefore this aspect is particularly interesting also from epistemological and historical point of view in order to trace the genealogy of particular approaches to the urban metabolism that are still being used and also to identify concepts that underlie particular methods like for instance energy analysis and more specifically concepts of agency that could be defined and extracted from studies using such methods and to conclude so our analysis identify actors that are inter-component subsystems and sub-processes of urban sociopolitical systems like for instance the case of network 4 which is particularly representative or structural features of the analytical and normative contexts that allow identifying such systems the urban metabolism and related characterization of agency that provided by the conceptual framework that presented at the beginning of this talk that entails that urban sustainability depends on the interplay between different forms of agency that can be ascribed to elements constitutive of urban sociopolitical systems and cells in this sense our conceptual framework sheds light on the urban metabolism notions potential as a tool for understanding that achieving urban sustainability by foregrounding the agency of structures and processes that comprise urban sociopolitical systems and those conclusion have to be however this conclusion have to be taken with some caveats because for instance they reflect the quantitative bias towards industrial ecology that is present in our source documents second the result of our analysis depend on the effectiveness of the pre-processing techniques that we chose to treat the documents before the analysis and third the relations mapped by the semantic networks planted in paper are predicated on a correct estimation of the correlation between the different aspects of the pattern of the distribution patterns of the board we consider and this implies using selected and most suitable metrics for instance Pearson correlation coefficient as an alternative to the Spearman correlation coefficient or the Kandall style correlation coefficient which can often cause for debate and since I have been talking for far too long measurements those my co-author for this document for this mouse fit Elisabetta and Ezen who are working with me on the different aspects of this project on the deal with the organ metabolism from the conceptual side and from the modeling side no questions Thank you for the presentation it was super clear at least the metallology aspects, the metallological aspects were super organized and clear and replicable. I was just one technical question when you were searching for the terms in the articles was it relevant where the terms were placed in the article like in the corpus of the article in the introduction in the result or near a verb because when you are doing text mining normally you can pay attention to that just just wondering Thank you Transfer question we we selected for each article we selected title we extracted the title the abstract and associated keywords so we only consider those parts of those sections of the article and as for the second part your question the relative position of words is captured by the use of engrams so by introducing multiple expressions we tried to take into account the context of different words and this we opted for the solution because it was conducive to our goal which was to identify to map of the semantic field for the expression urban metabolism so we wanted to be able to perform a keyboard analysis with multiple expressions and the use of engrams to capture multiple expressions was also a way to introduce context relevant context into this keyboard analysis so you didn't screen the whole text I used to continue because if you said in there were 24 articles the sources the sources I had to scan I guess that if you want to span it to the whole text it would be a large larger matrix but anyhow if you then you create engrams and then you do a preselection of the most frequent engrams using the enclosures that you have frequency and position but doesn't that come from the current matrix where you have to put every single one of the engrams together and then the documents and just it's a curiosity of mine because I wanted to understand sorry first it was an extension of engrams if you use the whole text the number of rows that you use in a matrix you kind of define it depending on the engrams you decide to select but and if you use the text you will still have an arbitrary decision on how the matrix could be but also this decision of how many engrams use them I wanted to understand is defined you use a current matrix that's kind of what I mean that's the basis the scale of this matrix depends on the amount of engrams you decide to choose so like if you use the whole text you would have a more large amount of data to start from to create your current matrix because from the keywords at least in a way it's good I use also keywords that if you want to turn into the stuff that they were have you thought of it or has competition actually in some sorts any question actually so we had to we were faced with tradeoff between completeness and as a computational cost associated with analyzing whole articles or using or trying to or introducing a higher let's say higher order engrams and therefore we first we decide to use for each document title, abstract and keywords because it seems to has a good compromise between only using the abstract or using the whole document and because in this way we can we could let's say we start from the assumption that abstract since there is a very tight word count limit limitation tight word counts that are applied to them we start from the assumption that abstract can be regarded as containing let's say words that are whose weight is important for at least from the point of view of those who wrote the authors of the article then we decide to expand to expand albeit in a limited way by including the title and the keywords starting from a similar assumption i.e. from the assumption that sometimes author they choose the title which words to use in the title and the keywords with this complementary respect to the words that they chose to use in the abstract itself and also to answer your question with respect to the Simpson coefficient so basically the Simpson coefficient is a set theoretical measure that basically corresponds to the ratio between the size of the subset of documents in which which terms i and j occur divided by the size of the smallest subset of documents between the subset of documents between the subset of all the document contains word term i and the subset of all documents that contain word j No, not yet, I'll let you know Stein had to go, he says he has questions, he's going to find you the next time he sees you but he had to go so so I want to kind of come back to the big picture for a little bit so what I take it from the context that part of what you're hoping I'm trying to get at what you're hoping that digital analyses are going to give you so I take it that part of this is you're hoping that you have competitors in the literature who don't see the agency stuff as being important at all and so you're hoping to be able to use this to sort of bash them over the head with the idea that like, look if you go to the literature and you look at the papers and you see what's important to the papers and then you ask yourself the question does what's important to the papers have to do with agency the answer is yes I guess how should I put it at that general level are you happy with the large hammer that you built do you think that you got a powerful enough hammer to convince the people who don't agree with you and if if you're not, why aren't you and how do you think you could build a better hammer I want to get to that, head that way for a second yeah I think that so the digital analyses must have full which provide a sort of basis that we could use as something point for interpreting the results but we we also used interpretive framework that relies on a particular take on the literature on agency and on metabolism that has been produced within a particular context, in a different context normal vertical ecology and our idea was to try to use this draw draw interesting conclusions from a particular disciplinary field and see how those conclusions or indications could apply to a very different field but very different which deals with similar phenomena and in this we wanted to test the footfulness of an interdisciplinary approach albeit limited because within the broader field of ecology as applied to field of ecology as the study of interaction between society and nature and the other problem or limitation of the digital approach is that we came up with association between terms with documents from selected that can be used to get more precise context about how those words are used and about the meaning of those words but the gap between words and concepts the process of extracting concepts from words in context requires additional additional theoretical more theoretical contributions and that's why we are working also at a more conceptual level let's say for instance I have been delving into odums howardolums and hpe odums partly based on the results of our computational analysis and interpretation we developed through the concept that the interpretive framework I presented today and Izabeta is working on other strains working on Niklas Luhman by authors belong to the social ecology approach the reference briefly would you like to see something Izabeta not really kind of question but besides this I hope I answer the question maybe it's related but I'm not sure I'm right in what I'm saying yeah I mean agency is I was trying to reflect on how this summer if we or they are happy with this summer and how far can we go with this summer somewhere I was thinking about the fact that in our political ecology agency is used as you mentioned to underline power dynamics beyond and behind and over everywhere around our metabolism and resources flows and processes within the city and I was wondering whether and this is also maybe linked to the farthest that we are taking with Nikola a more kind of more conceptual analysis to grasp concept and signification of agency and to maybe unpack also the political and the politics let's say more than the political power dynamics beyond embodied in agency itself so I guess maybe I don't guess I hope this summer could be yeah not I mean it's first to show agency is there and it's important to grasp agency because it's then also to connect with the politics of argumentalism of flows and resources to unpack what resource flows and processes their effects and also how that was not that clear and trying to reflect it was not really a question it was just something that I wanted to share maybe the question could be how much from to what extent from the lexical analysis can grasp values and politics and be embodied in agency in our agency is used because I think that based on the methods and tools we have used so far we can have our first sense of also of those normative aspects related to the agency of capitalism of urban society social systems to go how to go to go forward further in that direction I think it would be useful to use other tools more let's say more oriented toward the qualitative aspects qualitative aspects of text we are analyzing something in the vein of a sort of sentiment analysis but this would require an effort to categorize terms that we can associate with values so this would the way you see apply first attempt to try to map the axiological teams before the landscape of environmentalism no, I got the idea I have two precision questions after that one more substantial so you can cut me so the precision questions so you say you did some relative frequency analysis I didn't see anything in the results so did you find something? actually the problem with the relative frequency is that it ended up not influencing the average correlation coefficient because basically the documents in the concern is though because basically the years we were considered were only 30 so the variance was very limited can I? my other question is I missed how the disciplinary field entered in the data so the disciplinary fields that you were describing in the beginning they are associated in the data to sources to papers to keywords I missed how do we enter in the we assign each with a disciplinary category so for instance for social ecology we assign we looked at title, authors journal source document reference in the paper content of the paper we did this in Spain but it is not the discipline coming from the journal from the like the Scotland yeah we also use Scotland but sometimes the result were not convinced that information given that we set from Scotland was not accurate enough from our point of view okay so you did yeah and then we tried to map this the disciplinary affiliation of the different papers on the terms contained in the papers and mine third question so I was quite maybe it's because I missed the first seminar I was confused most of you are talking not about the techniques but about the connection between the questions at the beginning and the means so now I understand the goal is not necessarily to directly learn about the GenC and urban metabolism is to have a hammer against another group that is defending something else which now I see okay because for me it's quite far usually when we do data mining like that in in scholar paper is to learn about the history of science or the history of concepts inside the academia it's not to study stuff in the world directly most of the time but it's very important to learn how things change in academia because these are all academy papers so there's a part of the project about the GenC itself you know that tool style stuff or other and there's a part about using that kind of data to argue about how we talk about it or how we conceptualize it in these different sub fields I see the aim is to highlight and untap the potential which is the potential of using the concept of agency to clarify existing refined existing methodologies and identify other possibly define other methodologies in the field of raw metabolism studies that could address questions related to how flows of resources are studied and governed so there is an epistemological aspect and there is a normative particular aspect and the concept of agency is interesting because it allows to travel between those two fields those two aspects so the aspect related to the question of how are processes related to those flows of energy and matter that go through it is studied how are they understood and who produces knowledge about those processes on the one hand on the other hand there is the question of how those flows of matter and energy are oriented because they are they underpin how a city develops as a social system and how resources are distributed within this particular social system ok I understand better I'm quite skeptical but I understand better When you're doing the cluster analysis I see that you present in the table that you for each cluster you are assigning ten different studies and all of the clusters have the same amount of papers, right? Out to them Out to them but these papers or these studies are representative of the cluster because the words the keywords are appearing most frequently in those papers Are you planning to do like an in that review of these ten papers in order to try to find the way that the indicators are measured in order to construct your framework Are those the next steps? Thank you In the framework of this cluster we are mainly focusing on agency and the question of agency so we are selecting the papers based on how relevant are to the interpretive framework where we establish interpretation and that could be an interesting endeavor but let's say it exceeds the scope of our project So the framework that you are planning to build is only going to be conceptual, right? It's not going to be like you cannot measure of the indicators and then see what measures implied in this kind of measure of the indicator Actually the final outcome would be conceptual framework characterizing different aspect of the agency of the euro metabolism and in connection with the model classification system on which Evan is working and the connection between those two outcomes would be a model decision system relating different kind of models to different kind of actors in different forms of agency So that the actor can choose the better model to Yeah exactly I just wanted to ask that this concept of agency or any word related with agency you used it in the analysis part of all the papers that you did not use it it was related with urban metabolism and then from the concept of agency the results of the analysis Yeah basically we wanted to look for traces of the concept of agency beyond the term sorry but you did not in the analysis itself in creating the papers and in that part of your work you did not use any word specific to agency and in the normal the selection was centered around urban metabolism so we selected but centered around urban metabolism and the fields the system based field approaches constituted by urban ecology, industrial ecology and social ecology so we selected the first the 24 source documents based on this criteria so documents that were either dealing with urban metabolism or with the conceptual foundations of industrial ecology urban ecology and social ecology and then we gathered the papers, the articles that went into the final corpus based on the reference on the references of those 24 source documents Okay Like if you have these four networks so I have no experience in this kind of very little I have no experience in this kind of analysis so if you wanted to verify your results again by doing a similar search to see that whether you are the concept of agency or actors and other agencies like is it validated or there is no way of validating it quantitatively or maybe you can explore that or Yeah there are different ways for instance there are indirect ways like but with the let's say the interpretation do you mean the interpretation agency or the validity of the clustering Yeah, I mean the validity of the clusters Like maybe you are doing it first in trust to be my anchor but there are very direct ways for instance like starting from the assumption that if the clustering is meaningful the documents can assign to each cluster must be homogeneous from a semantic point of view we can use measure to like we can test for this homogeneity there are measures to be used and this could be one way of validating indirectly validating the trust to be loving algorithm Have you compared to for a buzz viewer clustering analysis No, we are not Yeah, so there is the qualitative validation like some expert from the field just feeling that the words are meaningful they are right And in this field usually you cannot validate it by using two search engines and then again in Web of Science maybe to see whether it is a kind of thing You could check by starting from another list of 20 papers Yeah You would work with someone from Gaia Leven if you want Web of Science Yeah But that's another word Yeah I have You are about to become more popular Yeah But usually that is not done here it is mostly the measures which we are saying of homogeneity and things like that which are used both by validators Yeah this approach was tailored to our to our goal so we assembled different bits and pieces of various approaches Yeah, I was going to say at some point I mean this would be to to far field for now but I want to talk to you about how you basically where you found all your legos and how you decided to put them together because it is a cool collection of legos but I mean it is I know you might be looking away with writing a methodology paper about this if this really is I am not sure that I have seen this pile of legos assembled this way before which is really cool Yeah I should identify like robust robust and not too time consuming way of auditing for a methodology paper Yeah I mean should that be the only thing I think you would need to be able to get a publication out of that side of this because I think it is Yeah It is easy to say that it is a robust and careful method of validating the results The main bias comes from the sample of the techniques so we should do it many times and check and do some statistical analysis to check that the method that you use is pretty robust and that is a lot of work but you could do it it is a paper it is a paper inside two metrics And this is where your skepticism comes from No My skepticism comes from the thing you want the method you use to get to what you want And I think what you want and the method is not exactly but I say the same thing when we have a new tool we want to use it but sometimes we forget what was the question we wanted to answer so maybe you should tailor the question to the tool but it seems that the question is much more ambitious than what you can do with that method So but he answered well he explained that it is a part of the process No, of course as I am working with him But I am skeptical because these techniques are powerful but not as much as we would like So we have to tailor the question very well to say the constant fight I can answer with that method Exactly Because at the end your cluster is done I am sure they are well done maybe we came out of the 24 first papers but let's say you had a good sample you have this cluster that seems okay it is very weird you seem from someone in the fields there is very strange words there your analysis through the paper is very clear and very extensive during the talk but how is it related to how is it exactly what you need to answer your question about agency I don't know So in your answer just because you will start again in your answer what I thought was interesting is that you say I concentrate on the producer of knowledge about agency and not on agency that's clever it's a good way to connect the method with the subject but if you are more interested by politics of agency maybe another method will be better it's been definitely more than time we can freely go thanks