In the context of Systems Thinking and Systems Dynamics, methodologies are used to understand and analyze the dynamics of complex systems. Specifically, Causal Loops Diagrams (CLDs) are a key tool for visualizing causal relationships and feedback of such systems. Thanks to adopting a specific metamodel, this thesis introduces a data-driven approach to explore them and enhance support for decision-making processes. The main objective is to develop a tool that automates the extraction and analysis of information contained in CLDs, currently performed manually with a laborious and time-consuming process. The approach adopted involves the representation of CLDs within a graph database, which is used to extract properties through targeted queries. The structured results of this first analysis phase are stored in a relational database and then made accessible through the "CLD-Explorer" application, developed using a low-code platform. "CLD-Explorer" offers a user-friendly interface that allows users to select, view, and analyze CLDs in just a few clicks. Through navigation and dynamic filters, the application allows the exploration of CLDs’ relevant aspects such as variables, relationships, feedback loops, and causal routes. To exemplify our approach, we applied it to three causal loop diagrams representing real-world systems adopted in health, economic, or social contexts, i.e., regarding the management of measures during the COVID-19 pandemic or the analysis of the fashion industry’s impact – here, we demonstrated the usefulness of CLDs in identifying areas for intervention. Overall, this work enforces how the automation of CLD analysis can effectively replace the current manual approach, reducing time and operational complexity, thus opening up new possibilities for more structured and data-driven complex systems and providing a significant contribution to decision-making processes.
Nell’ambito del Systems Thinking e Systems Dynamics, vengono utilizzate metodologie per comprendere e analizzare le dinamiche dei sistemi complessi. In particolare, i Causal Loops Diagrams (CLDs) rappresentano uno strumento chiave per visualizzare le relazioni causali e i circuiti di tali sistemi. Grazie all’adozione di un metamodello specifico, questa tesi introduce un approccio basato sui dati per esplorarli e migliorare il supporto ai processi decisionali. L’obiettivo principale è lo sviluppo di un tool che automatizzi le operazioni di estrazione e analisi delle informazioni contenute nei CLD, attualmente svolte manualmente con un processo laborioso e dispendioso in termini di tempo. L’approccio adottato prevede la rappresentazione dei CLD all’interno di un database a grafo, utilizzato per estrarre le proprietà tramire query mirate. I risultati strutturati di questa prima fase dell’analisi vengono memorizzati in un database relazionale e successivamente resi ac- cessibili tramite l’applicazione "CLD-Explorer", sviluppata utilizzando una piattaforma low-code. "CLD-Explorer" offre un’interfaccia user-friendly che consente agli utenti di selezionare, visualizzare e analizzare i CLD con pochi clic. Attraverso funzionalità di navigazione e filtri dinamici, l’applicazione permette di esplorare aspetti come variabili, relazioni, circuiti e percorsi causali. Per esemplificare il nostro approccio, lo abbiamo applicato a tre Causal Loop Diagrams che rappresentano sistemi reali adottati in contesti sanitari, economici o sociali come la gestione delle misure durante la pandemia di COVID-19 o l’analisi dell’impatto dell’industria della moda – qui, abbiamo dimostrato l’utilità dei CLD nell’identificare le aree di intervento. In generale, questo lavoro sottolinea come l’automazione dell’analisi dei CLD possa sostituire efficacemente l’approccio manuale, riducendo il tempo e la complessità operativa, aprendo così nuove possibilità per sistemi complessi più strutturati e basati sui dati e fornendo un contributo significativo ai processi decisionali.
A data-driven approach for storing and analyzing Causal Loop Diagrams
MAFTEI, LAURA DANIELA
2024/2025
Abstract
In the context of Systems Thinking and Systems Dynamics, methodologies are used to understand and analyze the dynamics of complex systems. Specifically, Causal Loops Diagrams (CLDs) are a key tool for visualizing causal relationships and feedback of such systems. Thanks to adopting a specific metamodel, this thesis introduces a data-driven approach to explore them and enhance support for decision-making processes. The main objective is to develop a tool that automates the extraction and analysis of information contained in CLDs, currently performed manually with a laborious and time-consuming process. The approach adopted involves the representation of CLDs within a graph database, which is used to extract properties through targeted queries. The structured results of this first analysis phase are stored in a relational database and then made accessible through the "CLD-Explorer" application, developed using a low-code platform. "CLD-Explorer" offers a user-friendly interface that allows users to select, view, and analyze CLDs in just a few clicks. Through navigation and dynamic filters, the application allows the exploration of CLDs’ relevant aspects such as variables, relationships, feedback loops, and causal routes. To exemplify our approach, we applied it to three causal loop diagrams representing real-world systems adopted in health, economic, or social contexts, i.e., regarding the management of measures during the COVID-19 pandemic or the analysis of the fashion industry’s impact – here, we demonstrated the usefulness of CLDs in identifying areas for intervention. Overall, this work enforces how the automation of CLD analysis can effectively replace the current manual approach, reducing time and operational complexity, thus opening up new possibilities for more structured and data-driven complex systems and providing a significant contribution to decision-making processes.File | Dimensione | Formato | |
---|---|---|---|
2024_12_Maftei_Tesi.pdf
accessibile in internet per tutti a partire dal 18/11/2025
Descrizione: testo tesi
Dimensione
7.65 MB
Formato
Adobe PDF
|
7.65 MB | Adobe PDF | Visualizza/Apri |
2024_12_Maftei_Executive Summary.pdf
accessibile in internet per tutti a partire dal 18/11/2025
Descrizione: executive summary
Dimensione
943.23 kB
Formato
Adobe PDF
|
943.23 kB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/230403