Demo project built by ZOHREEN to showcase our work.
Synthesizing high-density Mapbox vector rendering, custom binary SQLite tile distribution, and an ultra-low battery GPS throttling protocol to navigate historically dense, connectivity-compromised cultural corridors without losing a single frame.
Autonomous geometric vector pipeline parsing protobuf payloads straight from local storage.
Vector engine profile recorded across mid-tier low-spec hardware under heavy raster stress.
Karachi presents one of the most demanding navigational environments on earth. Intense high-density concrete streetscapes create acute cellular dead-zones; handoffs between 4G and 2G towers routinely stall active HTTP connections. For urban explorers, historic researchers, and visitors navigating the serpentine bazaars of Kharadar and Old Clifton, conventional consumer mapping applications failed completely.
Standard mapping platforms relied on continuous online vector tile requests over the air, exhausting device batteries within three hours of heavy walking tour use. Furthermore, cultural landmarks with high historic gravity often lacked standardized postal addresses, demanding a precise sub-meter spatial coordinate index.
Over 62% of historic pedestrian alleys lack steady 3G data throughput.
Continuous HTTP map streaming caused extreme chipset heat in 38°C ambient climate.
Rather than requesting raster map tiles over HTTP, we compiled vector geometry into a compressed binary SQLite database embedded on-device. We developed a proprietary low-level deserializer that reads Mapbox vector tile (MVT) protocol buffers directly from localized flash memory.
Bloc architecture synced to an embedded Isar NoSQL reactive store with sub-5ms UI reconciliation.
Mapbox GL native bindings directly interfacing with Metal (iOS) and Vulkan (Android) pipelines.
Intelligent differential sync applying lightweight JSON patch changes exclusively during verified Wi-Fi sessions.
The production application merges extreme technical discipline with luxury Arabian dusk aesthetics. CitiGuide features an intelligent dual-mode UI that transitions smoothly between maritime dark mode and high-contrast ivory daylight display, accompanied by localized bilingual voice-guided walking tours.
Culturally nuanced voice guides recorded natively in Urdu (اردو) and English.
Sensor polling drops from 1Hz to 0.1Hz when accelerometer registers stationary standing.
Midnight Teal for twilight tours, Ivory Sunlight for direct tropical noonday glare.
Native rendering performance across iOS and Android chipsets under active pinch-zoom stress.
Curated walking itineraries and architectural bookmarks cached locally with zero network connection.
Unified Dart codebase compiling to native ARM performance across iOS and Android.