Dynamic Map-based Data-Centric Approach for Tourism and Cultural Heritage Preservation Digital Twins

Abstract: Tourism is an essential and growing economic activity worldwide, bringing benefits such as job creation, revenue generation, and tax revenue, and driving economic prosperity. Tourism activities may also have negative impacts on cities, including overtourism and pressure on housing and real estate markets. Cultural heritage is an essential asset of cities and countries that must be preserved. Cultural heritage, as an asset, is commonly explored through tourism activities and may have negative impacts, including physical degradation, commodification, and loss of authenticity. A digital twin is a data-driven virtual representation of a physical object, system, or environment. However, although digital twin technology has been widely adopted in manufacturing, Industry 4.0, and urban planning for smart cities, there remains a gap in specialized digital twins for tourism and cultural heritage management. This paper proposes a QGIS-based, data-centric approach to digital twin frameworks that supports the management, development, and deployment of tourism and cultural heritage services and applications in smart cities. The data-centric approach is embedded in a specialized digital twin focusing on the Salvador Historic Center - Pelourinho, a highly important cultural asset and tourism spot for the city of Salvador. Currently, Pelourinho faces a persistent challenge in sustaining tourism flux while safeguarding its cultural and heritage assets. Preliminary results indicate that the data-centric approach adopted by Pelourinho's DT facilitates data visualization, integrates data silos, and adequately supports management, enabling managers to address heritage preservation and conservation issues, control over-tourism, and implement urban resilience and climate adaptation measures.

• The paper proposes a data-centric digital twin that integrates IoT sensors, heritage inventories, social data, and geospatial records through PostgreSQL/PostGIS and QGIS.
• The architecture combines data harmonization, spatial analytics, tourism-flow estimation, microclimate and sound models, and dynamic maps to support congestion monitoring and what-if planning.
• The framework offers a coherent design for heritage and tourism management, but it still requires live deployment to validate accuracy, real-time performance, privacy safeguards, and institutional data sharing.

Context and motivation

Tourism is a major economic sector: 1.4 billion international arrivals in 2024 generated $10.9 trillion in economic activity (about 10% of global GDP) and 357 million jobs, roughly 10.6% of the global workforce (2603.00079). Yet tourism pressure on historic urban cores produces overtourism, physical degradation of heritage assets, commodification, and loss of authenticity. The paper addresses this tension for the Historic Center of Salvador, Brazil — Pelourinho, a UNESCO World Heritage Site — which suffers from unmanaged tourist flows, urban heat islands, noise pollution, high building density, scarce green space, and fragmented institutional data held by bodies such as IPAC and IPHAN.

The authors identify a threefold research gap: existing digital twins rarely treat tourism, cultural heritage, and social issues concomitantly; heritage DTs are mostly digitization/visualization tools lacking data-rich models; and few DTs embed AI over multi-domain smart-city datasets. The central research question is whether a data-rich strategy can be integrated into a specialized DT framework that holistically manages tourism and cultural heritage.

A QGIS-based, data-centric digital twin

The proposal is a specialized DT for Pelourinho built around a data-centric design: rather than prioritizing advanced data models, it emphasizes improving the quantity and quality of data, motivated by the shift toward data-centric AI (2603.00079). The core technical choice is PostgreSQL with the PostGIS extension, selected for its native spatial capabilities (2D–4D, spatial relationships, analytics) and scalability, coupled with QGIS as the georeferenced integration and visualization layer. Since QGIS does not natively support dynamic data-centric operation, the system employs a database-to-QGIS mapping in which database tables and views mirror the QGIS layer structure, ensuring time-series mappings and relations.

The database comprises four logically distinct but integrated stores: a geographic QGIS database; an IoT database of time-referenced sensor parameters; a heritage/historical and social data store; and an application-defined database supporting customization to different management strategies. A Service-Specific Data Processing and Analytics (SSDPA) module performs pre-processing directly in the database via spatial SQL queries and views before results reach the visualization layer.

Architecture

The DT architecture is organized hierarchically into three main blocks plus an application tier:

• Physical layer: sensors capturing temperature, humidity, noise, people flow, and microclimatic conditions, complemented by institutional sources (urban registry, IPAC/IPHAN heritage inventories), 2D/3D models, satellite imagery, meteorological data, and public geospatial databases.
• Data Hub: the structural axis of the architecture, comprising a raw data zone, a processing and harmonization pipeline (noise removal, inconsistency correction, temporal synchronization, QGIS-based georeferencing, multi-source merging for density estimation), and structured repositories (geographic, time-series, and application-specific stores).
• Analytical models and simulation core: human flow, microclimate, and sound-propagation models producing maps, indicators, forecasts, and simulated scenarios, with predictive analytics and continuous updating.
• Visualization and applications: dynamic maps, dashboards, and management interfaces.

The authors argue that the data-centric strategy is not an add-on but the element that articulates all components, mediating between fragmented real-world data and models that require consistent inputs.

Tourism flux use case

A use case formalizes tourist-flow management with three actors — Urban Manager (public authority), Tourism Planner, and Researcher/Analyst. The core function, monitoring tourist flow, includes estimating tourist density via fusion of camera detections, passive WiFi/Bluetooth signals, and statistical estimates, and identifying congestion hotspots. Complementary use cases cover "what-if" scenario simulation (events, route changes, seasonality), assessment of environmental impacts by correlating density with acoustic pressure and temperature, and aggregated decision support. This predictive capacity distinguishes the DT from conventional monitoring systems.

Results and claims

The paper reports preliminary, qualitative results rather than quantitative benchmarks. It claims that the data-centric approach facilitates data visualization, integrates previously siloed datasets across institutions, and adequately supports managers in addressing heritage preservation, overtourism control, and urban resilience/climate adaptation. These claims should be read as design-level validation: no deployment metrics (e.g., density-estimation accuracy, latency, or measured reductions in overcrowding) are presented, so the adequacy assertion rests on architectural argumentation rather than empirical evaluation.

Limitations and open questions

Several limitations are implicit or conceded. The evaluation is preliminary and lacks operational evidence from live deployment in Pelourinho. The framework presupposes cooperation among multiple institutions whose data fragmentation is itself cited as a structural obstacle, yet governance mechanisms for data sharing are not specified. Privacy implications of camera and WiFi/Bluetooth tracking of visitors are not addressed. The AI component is asserted as pivotal but the specific algorithms, their training data requirements, and validation remain future work. Open questions include how accurately heterogeneous fusion (cameras, passive signals, statistical estimates) can estimate density in dense historic streetscapes, and whether the PostGIS/QGIS stack sustains real-time performance at full sensor scale.

Conclusion

The paper contributes a replicable, QGIS map-based, data-centric DT framework for joint tourism and cultural heritage management, instantiated for Salvador's Pelourinho Historic Center. Its principal strength is architectural coherence: a centralized spatial database with database-to-QGIS layer mapping that integrates IoT time series, heritage inventories, and social data under one georeferenced structure. Future work targets concrete services such as urban heat-island modeling, real-time visitor flow monitoring, dynamic route recommendation, façade degradation forecasting from heat stress, restoration-scenario simulation, and green-area planning. Empirical validation of these services in operation remains the key outstanding step.