This dataset contains a digital urban scenario, named Tomorrowville, that is developed as a testbed for multi-hazard risk assessments and to evaluate the performance of urbanisation scenarios. Tomorrowville was created to represent a global-south urban setting by means of its socio-economic and physical aspects. It covers an area of 500ha located south of Kathmandu (Nepal). The dataset consists of 5 different data types: - Buildings: Data representing the building footprints for today and 50 years from now including specific attributes to be used within multi-hazard risk assessments. - Land uses: Data representing the land use information for today and 50 years from now. - Vulnerability: Tabular files that contain vulnerability functions for buildings under earthquake and flood hazards. - Household: Data that contains social attributes of the Tomorrowville, such as the level of education, age, gender and working status of the individuals and their states in the households. - Hazards: Data representing the hazards (earthquake (eq), floods (fl) and debris flows (df) that may impact the case study areas of Tomorrowville. Observational data of the built environment and socio-economical properties of Kathmandu and Nairobi were used in addition to synthetic social data to create the initial scenario. This is a synthetic social dataset, meaning it was derived from existing population projections and distributions for the testbed but does not reflect the reality on the ground. It is synthetically created using specific algorithms in a GIS environment to represent a Global South social context. For the building data, Open Street Map (OSM) database is used as a basis. The data is scraped from OSM and modified to represent an urban context for Tomorrowville. The attributes are also modified to be able to use in a multi-hazard risk computation. A taxonomy string is generated for each building that represents an acronym for its building code level, number of storeys, occupation type and structural system. The hazards that were existing in the selected spatial extent were earthquake, flood, and debris flow. Hazard data represents an intensity measure for the relevant hazard type (ground acceleration for earthquake, flow velocity for the flood and debris flow hazards). The following hazard input data are included: - For the flood simulations, the discharge and rainfall time series are generated based on moderate to peak daily data based on recorded data from the Department of Hydrology and Meteorology, Nepal. - Earthquake hazard sources are generated and simulated by Jenkins et al. (2023). - For the debris-flow and flood simulations tri-stereo Pleiades satellite imagery is used to produce a 2m resolution Digital Elevation Model. The work to create this dataset is supported by NERC as part of the GCRF Urban Disaster Risk Hub (NE/S009000/1) Full details about this dataset can be found at https://doi.org/10.5285/8b5834a5-ae8a-4f24-836c-16fab961aeb3