<![CDATA[Urban heat exposure in tropical megacities has intensified, yet public weather services still communicate conditions as separate meteorological parameters that do not describe the thermal load experienced by the human body. The Physiological Equivalent Temperature (PET) index addresses this limitation, but most PET studies conclude as analyses, maps, or datasets rather than services reaching the public. This study designs and evaluates a multichannel information architecture for delivering PET-based thermal comfort information at subdistrict resolution to the general public in DKI Jakarta. Hourly ECMWF Integrated Forecasting System forecasts were obtained through an open weather data service for 267 subdistricts across 44 districts and six administrative cities, from which mean radiant temperature and PET were computed and classified using a nine-class tropical scheme. The information was organised into three separated nodes in which 2,448 precomputed aggregate records accompany 12,816 hourly subdistrict records, delivered through an interactive dashboard and a scheduled messaging summary. Precomputation adds 9.3 percent to stored payload but reduces the data read to render the aggregate layers from 3,053.2 KB to 5.90 KB, a factor of 518. All twelve functional test scenarios performed as expected, and the mean usability score across twelve respondents reached 85.2, well above the reported average. Separating aggregation levels and pairing a pull channel with a push channel makes high resolution thermal comfort information navigable for non-expert users.]]>
