iot-data-sandbox

---

A Go-based IoT data sandbox that enables LLM-driven analysis and manipulation of time-series data.

Users interact with the system through natural language prompts:

• "show me any anomalies that occurred in the last 3 days"
• "plot the moving average with a window of 6 hours"

An agentic AI agent orchestrates a registered set of tools to fulfill these requests, ranging from simple statistical analysis to AI/ML forecasting models and anomaly detection.

Project Structure

Repository	Description
iot-data-sandbox	Go backend API server
iot-data-sandbox-ui	Vue.js frontend application

Quick Start

1. Start the backend

git clone https://github.com/nathanaday/iot-data-sandbox.git
cd iot-data-sandbox
go run cmd/server/main.go

The server runs on localhost:8080.

2. Start the frontend

git clone https://github.com/nathanaday/iot-data-sandbox-ui.git
cd iot-data-sandbox-ui
npm install && npm run dev

The frontend expects the backend to be running at localhost:8080.

Entity Model

The entity hierarchy follows: project -> datalayer -> datasource

• Project - Logical container for datalayers; enables save/load of entire workspaces
• Datalayer - Wraps a datasource with visual properties (color, theme, z-index) for UI display
• Datasource - Manages CSV data and tracks timeseries metadata (length, labels, time span)

Generate API Docs

go install github.com/swaggo/swag/cmd/swag@latest
swag init -g cmd/server/main.go

References

• langchaingo
• NA CO2 Emissions Dataset