Πλοήγηση ανά Επιβλέπων / ουσα "Dioikitopoulos, Evangelos"

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω

Τώρα δείχνει 1 - 6 από 6

Airbnb beyond stars: what reviews reveal
(2025-12-18) Kuqja, Franc; Pagratis, Spyros; Dendramis, Yiannis; Dioikitopoulos, Evangelos
Online reviews play a pivotal role in shaping consumer choices on platforms like Airbnb, where textual content offers deeper insights into guest experiences than numerical ratings alone, all while macroeconomic factors like GDP per capita, inflation, and unemployment shape expectations. This study analyzes Airbnb data from eight European cities, first linking pricing to property traits like capacity and ratings, host superhost status, and experience via regressions, then tracing sentiment from review texts over time, spatial city clusters, and economic conditions. NLP converts reviews into sentiment scores, balanced annually per city and aggregated at listing and city-year levels, with sigmoid mapping to Airbnb's 1-5 scale for direct rating comparisons. Token analysis pinpoints words driving positive/negative sentiment, revealing city-specific review tendencies. Key findings: Accommodation capacity emerges as the dominant, consistent price driver, while ratings and host factors show inconsistent impacts. Sentiment uncovers rating inflation—scores cluster at the high end—plus geographic disparities proving uneven guest experiences within cities. Pooled regressions with city fixed effects link economic shifts associatively (not causally) to sentiment fluctuations, though effects remain modest. Ultimately, fusing NLP sentiment analysis, spatial clustering, and macro indicators illuminates platform user dynamics, advancing platform economics and computational social science, despite challenges measuring sentiment in overwhelmingly positive review settings.
The role of culture and cultural and creative industries to regional innovation: evidence from European regions
(2026-03) Chatzianastasiou, Maria; Χατζηαναστασίου, Μαρία; Varthalitis, Petros; Kalyvitis, Sarantis; Dioikitopoulos, Evangelos
This dissertation examines the relationship between regional innovation and cultural employment, with a particular focus on the role of culture. While innovation was traditionally associated mainly with high-tech industries and scientific fields, recent research highlights that cultural and creative sectors also significantly contribute to economic growth through creativity and knowledge-intensive activities. At the same time, culture is recognized as a key factor shaping economic behavior and innovation capacity. The study has two main objectives: first, to analyze the relationship between cultural employment (used as a proxy for CCIs) and innovation performance across European regions; and second, to investigate the mechanisms through which cultural employment affects innovation, including R&D expenditures, patent applications, and human capital. Individual economic preferences are also used as proxies for cultural traits.
Behavioral determinants of retaliation: evidence from fouling behavior in European professional basketball
(2026-03-16) Papoutsis, Nektarios-Theodoros; Παπουτσής, Νεκτάριος-Θεόδωρος; Pagratis, Spyridon; Vrontos, Ioannis; Dioikitopoulos, Evangelos
This thesis examines whether professional basketball players in the EuroLeague exhibit retaliatory behavior through fouling (retaliatory fouling), that is, whether they tend to commit more fouls after previously being fouled, and to what extent this reaction is influenced by behavioral and cultural characteristics. The analysis combines player performance data from the EuroLeague with cross-country behavioral indicators, such as patience and negative reciprocity, drawn from the cultural preference dataset published in the Quarterly Journal of Economics. To empirically investigate these relationships, a series of linear regression models with robust statistical inference are estimated. The results support the existence of retaliatory behavior: the fouls a player receives are positively associated with the fouls they commit. Among the behavioral determinants, patience emerges as the strongest and most consistent factor. In performance models, higher levels of patience are associated with greater free-throw efficiency, a finding consistent with higher self-control and composure under pressure. Moreover, patience is associated with fewer fouls received, suggesting either a calmer style of play or better positioning that reduces physical contact. Negative reciprocity shows some interaction with defensive actions, such as steals, in predicting fouls committed. However, when retaliatory behavior is modeled directly and patience is included in the specification, the effect of negative reciprocity becomes weaker or statistically insignificant. These findings suggest that patience, rather than vengeful tendencies, is the primary mechanism limiting retaliatory behavior. Finally, life-cycle dynamics also play an important role. Age and career experience indicate that more experienced players commit slightly more fouls overall but react less intensely to provocation. EuroLeague veterans display significantly weaker retaliatory behavior, a result consistent with learning effects, strategic discipline, and improved emotional regulation. Overall, the findings indicate that retaliatory behavior is a real and measurable phenomenon in European professional basketball, but it is moderated by patience and career experience.
Economic and institutional determinants of vaccination uptake: a global panel data and machine learning approach
(2026-03-16) Rryci, Marilena; Pagratis, Spyridon; Dendramis, Yiannis; Dioikitopoulos, Evangelos
This thesis investigates the economic and institutional determinants of COVID-19 vaccination uptake using a global country-month panel dataset covering the period 2021–2022. Despite the rapid development and widespread availability of effective vaccines, vaccination coverage evolved unevenly across countries, indicating the presence of deeper structural and dynamic factors. The analysis combines high-frequency vaccination and epidemiological data with structural indicators of economic development and institutional quality. The empirical framework distinguishes between long-run structural determinants and short-run within-country behavioral responses. Two-Way Fixed Effects (TWFE) panel models are employed to identify dynamic effects, complemented by cross-country regressions. In addition, machine learning techniques are applied to assess predictive performance and explore potential nonlinear relationships. The results show that increases in COVID-19 cases and deaths are associated with statistically significant acceleration in vaccination rollout, reflecting behavioral responses to perceived epidemiological risk. Furthermore, economic development and government effectiveness emerge as key determinants positively influencing vaccination coverage, while income inequality exhibits a more complex and indirect role. Overall, the findings suggest that vaccination uptake is driven by the interaction between structural state capacity and dynamic behavioral responses. By integrating panel econometrics with machine learning methods, this study provides a comprehensive framework for understanding global disparities in vaccination outcomes.
Pricing, demand inference, and competitive dynamics in short-term rentals
(2026-03-16) Σολδάτος, Ιωάννης; Soldatos, Ioannis; Pagratis, Spyros; Tzavalis, Elias; Dioikitopoulos, Evangelos
This thesis examines how posted prices are formed in the short-term rental market in Athens when demand is not directly observed. A central methodological challenge is that public platform data provide information on prices and availability, but not on realized occupancy. To address this issue, the study uses repeated Inside Airbnb calendar snapshots for the period 2023–2025 and constructs a monthly listing-level panel. By separating non-market host blocks from actual market availability, computing a lower bound for booked nights, and estimating booking curves, the thesis develops a transparent and reproducible occupancy proxy that serves as its main measure of demand. Building on this panel, the study estimates posted-price models that combine structural listing characteristics, quality signals, seasonality, and a hierarchically defined peer environment, while also examining the within-listing relationship between relative price and occupancy over time. In addition, it develops diagnostic tools to assess potential underpricing or overpricing and to measure the intensity of dynamic pricing behavior. In a complementary empirical layer, the thesis uses operational reservation data from Airbnb and Booking.com to investigate how realized Average Daily Rate is associated with review propensity and guest ratings. Finally, in a policy-oriented layer, it examines the associations between short-term rental intensity and long-term rent and sale-price developments in Athens around the 2025 licensing intervention. The findings indicate that seasonality, location, and competition are major drivers of posted prices, that higher relative prices are associated with slightly lower occupancy, and that the policy-related results should be interpreted cautiously given the study’s identification constraints in causality.
Sentiment analysis on central bank speeches and economic policy
(2026-04-27) Skampardonis, Aristotelis; Σκαμπαρδώνης, Αριστοτέλης; Pagratis, Spyros; Tzavalis, Elias; Dioikitopoulos, Evangelos
The thesis explores the economic and behavioral consequences of central bank communication based on sophisticated text analysis methods. Whereas other studies have basically examined the financial market responses, this paper widens the scope of the study to denote the macroeconomic and structural response such as innovation performance and the national saving behavior. This study combines Natural Language Processing (NLP) techniques, all of which entail dictionary metrics, transformer-based models, and topic modelling, with panel econometric analysis by establishing sentiment, tone, and behavioral indices based on a cross-country corpus of speech by central banks. The empirical model is connected to communication determinants to the results of innovation (R&D expenditure), cultural traits (Hofstede dimensions), behavioral attributes (patience, risk aversion), and macroeconomic factors. Findings suggest that risk-oriented and warning communication has a negative relationship with the rate of investment and innovation, especially during crisis times, whereas long-term oriented language is positively related with saving rates. Communication tone is developed on a systematic level based on cultural dimensions, exposing cross-country policy narrative heterogeneity. Machine-learning models also indicate that behavioral constructs within speeches have predictive potential of macroeconomic performance, and the tone of caution has a greater forecasting ability of performance compared to the tone of confidence. In general, the results indicate that the communication of the central banks is one of the behaviors transmission channels that affect the expectations, intertemporal choices, and structural economic processes that the theorist of the monetary tools does not address.