Sub-district-level emergency call rates, response intervals, and spatial autocorrelation, with analysis code: Public Safety Center 119, Malang City, Indonesia (July–August 2025)
Description
This dataset accompanies a retrospective observational study of response intervals and geographic access in Public Safety Center 119 (PSC 119), the single-base municipal emergency medical services system of Malang City, East Java, Indonesia. It contains aggregated data for the city’s 57 sub-districts (kelurahan) within five districts (kecamatan), derived from all 222 eligible emergency calls recorded by PSC 119 from 1 July to 31 August 2025, together with the analysis code. Files: (1) Supplementary_Data_kelurahan_aggregated.csv – one row per sub-district: district and sub-district codes, sub-district name, number of calls, 2020 census population, call rate per 10,000 residents, median total response interval (minutes; blank where fewer than three calls), number of calls with a valid interval, local Moran’s I of Empirical Bayes-standardized call rates, pseudo P value (9,999 permutations), cluster type, and significance after false discovery rate correction. (2) Python scripts for the interval statistics, road-network service areas (NetworkX, Dijkstra’s algorithm on the 1:25,000 Rupabumi Indonesia road network), and figures. (3) A Jupyter notebook that replicates the spatial autocorrelation analysis with PySAL (esda, libpysal). Incident-level records are not included because they contain scene addresses; access is restricted under the ethics approval. Administrative boundaries and road networks come from public sources (Statistics Indonesia; Geospatial Information Agency) and are not redistributed.
Files
Steps to reproduce
Data generation: All emergency calls recorded in the Public Safety Center 119 (PSC 119) Malang City dispatch registry from 1 July to 31 August 2025 were screened. Non-emergency calls, classified independently by three investigators against the national PSC 119 emergency algorithm, and records with incomplete timestamps were excluded, leaving 222 calls. Scene addresses were geocoded with Google Maps and assigned to the city’s 57 sub-districts (kelurahan) by point-in-polygon overlay. Call counts were divided by 2020 census populations (Statistics Indonesia) to give rates per 10,000 residents. Median total response intervals were calculated for sub-districts with at least three calls. Local spatial autocorrelation was computed with local Moran’s I on Empirical Bayes-standardized rates, using first-order queen contiguity with row standardization, 9,999 permutations, and Benjamini–Hochberg false discovery rate correction. Reproducing the analysis: (1) Install Python 3 with pandas, numpy, libpysal, and esda (pip install pandas numpy libpysal esda). (2) Unzip the dataset and run code/reproduce/reproduce_spatial_autocorrelation.py. (3) The script reads data/psc119_malang_subdistrict_aggregated.csv and data/psc119_malang_subdistrict_queen.gal and reports the global Empirical Bayes-adjusted Moran’s I (expected 0.1808) and the local clusters (expected High-High: Bumiayu, Tlogowaru, Wonokoyo; Low-Low: Karangbesuki, Pisangcandi; Low-High: Kebonsari, Kiduldalem). Pseudo P values vary slightly with the random seed. The scripts in code/reference_scripts document the road-network and mapping procedures but require incident-level data, which are not shared (see README).
Institutions
- University of BrawijayaEast Java, Malang
Categories
Funders
- University of BrawijayaEast Java, MalangGrant ID: Rector Decree No. 2713 of 2025. Programme name: Globalizing UB Project