01 · ARTIFACT 003 / CASE STUDY
Back to selected workReal-timeIoTDashboard
Engineered during my Dashboard Development Internship at Flostat, this real-time monitoring interface visualizes sensor readings, flow telemetry, tank levels, and device states for IoT water management systems using React, TypeScript, MQTT, and REST APIs.
Role
Dashboard Development Intern (Flostat)
Category
REAL-TIME DATA / FRONTEND
Tech Stack
React · TypeScript · MQTT · REST APIs · Tailwind CSS · Data Visualization
Timeline
FLOSTAT INTERNSHIP · REAL-TIME SYSTEMS
02 · THE CHALLENGE
The problem.
BACKGROUND & CONTEXT
Industrial and municipal IoT water management systems generate continuous telemetry from distributed flow meters, pressure transducers, and level sensors. Operators need an immediate, legible interface to monitor live readings and detect anomalies.
TARGET USERS
Facility operators, water management engineers, and field technicians monitoring distributed pump stations and storage reservoirs.
CORE REQUIREMENTS
- Live sensor readings updated continuously via MQTT message streams.
- Historical and real-time data visualization for flow rate, tank capacity, and valve telemetry.
- Live monitoring interface with clear node health states and threshold alerts.
- Responsive dashboard layout usable on both control-room monitors and field tablets.
SYSTEM CONSTRAINTS
- High-frequency MQTT packets could trigger excessive React re-renders if state was not buffered carefully.
- Telemetry views had to remain scannable under varying ambient lighting conditions.
03 · THE APPROACH
The approach.
01 / RESEARCH & DISCOVERY
Studied operator workflows at Flostat to identify which telemetry metrics require instant glanceability (active flow rate, reservoir percentage, connection state) versus historical trend analysis.
02 / ARCHITECTURE PLANNING
Combined initial historical state hydration via REST API endpoints with a persistent MQTT WebSocket client that buffers incoming sensor payloads into a normalized state store.
03 / TECHNOLOGY SELECTION
Built with React and TypeScript for type-safe telemetry packet parsing, integrating MQTT client subscriptions alongside REST APIs and custom SVG/Canvas data visualizations.
04 / USER EXPERIENCE DECISIONS
Prioritized tabular numerals, high-contrast semantic threshold indicators (pairing explicit text labels with color), and a responsive grid that adapts cleanly from 1440px desktops to mobile field devices.
04 · SYSTEM ARCHITECTURE
Under the surface.
IoT Sensor Nodes & MQTT Broker
Distributed flow meters, level sensors, and pump controllers publishing telemetry topics.
MQTT Broker · WebSockets
REST & Stream Ingestion Layer
Combines REST endpoints for historical logs with live MQTT topic subscriptions.
REST API · MQTT Client
Buffered Telemetry Dashboard
TypeScript state buffer, real-time waveform visualizers, and responsive node cards.
React · TypeScript · Tailwind CSS
05 · KEY FEATURES
Feature showcase.
Live MQTT Sensor Readings & Telemetry Stream
Subscribes to hierarchical MQTT sensor topics to display live flow rates (L/min), reservoir levels, and pressure readings with sub-second visual updates.
MQTT LIVE STREAM · TABULAR METRICS
Multi-Node Water System Visualization
Interactive topology and time-series visualization correlating pump status, valve actuation, and tank consumption across connected water management zones.
TOPOLOGY MAP · RESPONSIVE GRID
06 · ENGINEERING CHALLENGES
The hard part.
Preventing UI Jank During High-Frequency MQTT Bursts
PROBLEM
Multiple IoT sensor nodes publishing telemetry simultaneously caused rapid React state updates and chart re-draws.
ENGINEERING SOLUTION
Implemented a ref-backed message buffer that coalesces incoming MQTT packets and flushes normalized state updates on a controlled cadence.
Resilient Reconnection & Stale-Data Signaling
PROBLEM
Field networks occasionally drop WebSocket connections, risking operators acting on stale sensor readings.
ENGINEERING SOLUTION
Added per-node heartbeat tracking and explicit connection state labels that visually flag stale telemetry and automatically reconcile with REST snapshots upon reconnect.
07 · RESULTS AND LEARNINGS
What came to life.
VERIFIED DELIVERABLES
- Production-grade React and TypeScript dashboard interface developed during the Flostat internship.
- Integrated MQTT live telemetry streaming alongside REST API historical data fetching.
- Responsive multi-panel layout supporting sensor readings, data visualization, and live monitoring.
KEY ARCHITECTURAL LEARNINGS
- Decoupling network ingestion frequency from UI render frequency is essential for smooth real-time dashboards.
- Clear typographic hierarchy and explicit status labels matter far more than decorative gauges in operational software.