HydroTemp OS – RPI Sensor Dashboard

A Shiny application for monitoring temperature and fan-control data logged by a Raspberry Pi at FlatBroke Farms. The dashboard ingests raw sensor logs, cleans them, and renders an interactive time-series chart so you can explore soil, ambient, enclosure, and on-board RPI temperatures side-by-side.

---

What It Does

Feature	Details
Multi-sensor time-series plot	Displays up to six data streams on a single chart with distinct, named colours
Sensor toggles	Sidebar checkboxes let you show/hide any individual sensor without reloading data
Date-range slider	Filter the chart to any sub-period within the available data
Fan-ON overlay	The "Fan Status" trace highlights moments the cooling fan was active by plotting RPI temp only while the fan is running
Automatic outlier removal	IQR-fencing (1.5 × IQR) is applied to noisy sensors before display

---

Sensors

Label	Source column	Colour	Description
Soil 4	Soil_4	Brown	Soil temperature at ~4-inch depth
Soil Surface	Soil_Surface	Green	Surface soil temperature
Ambient	Ambient	Red	Ambient air temperature
Enclosure	Enclosure	Blue	Temperature inside the RPI enclosure
RPI Temp	RPI_Temp	Deep pink	Raspberry Pi on-board CPU temperature
Fan Status	Fan_Status	Aquamarine	Fan-ON overlay (plots RPI temp only while fan is running)

---

Data Sources

Both files live in source_data/ and are read at app startup:

File	Content
OnBoard.log	Tab/space-delimited RPI log: date, time, CPU temperature, fan state (Fan: ON / Fan: OFF)
templog.csv	Timestamped CSV from the external sensor array: Soil Surface (Bad Sensor), Soil Temp 4, Enclosure, Ambient

The two datasets are joined on timestamp using a fuzzy (nearest-neighbour within ±4 minutes) inner join via the fuzzyjoin package, and all timestamps are converted to America/Chicago.

---

Project Structure

HydroTempOS/
├── app_rpi.R              # Main Shiny app
├── Setup.R                # Library loader & path definitions
├── RPI_data_processing.qmd  # Exploratory / pre-processing notebook
├── source_data/
│   ├── OnBoard.log        # RPI on-board log (date, time, CPU temp, fan)
│   └── templog.csv        # External sensor temperature log
└── .gitignore

---

Requirements

R packages

Package	Role
shiny	Web application framework
tidyverse	Data wrangling & ggplot2 visualisation
fuzzyjoin	Nearest-timestamp join between the two log files
lubridate	Timezone-aware datetime handling
here	Portable project-relative file paths
tidylog	Verbose join/filter feedback during development

Setup.R auto-installs any missing packages when the app is sourced. The here package auto-detects the project root from the nearest .Rproj, .git, or .here marker, so no manual root-finding call is needed.

Hardware context

The logs are generated by a Raspberry Pi running a Python-based sensor daemon that records DS18B20 temperature probes and a GPIO-controlled cooling fan. The Shiny app is purely a viewer — it does not connect to the Pi directly.

---

Running the App

1. Clone the repository and open the project in RStudio (a .Rproj file, .git directory, or .here marker at the repo root lets here resolve all paths automatically).
2. Ensure source_data/OnBoard.log and source_data/templog.csv are present.
3. Open app_rpi.R and click Run App, or execute from the console:

shiny::runApp("app_rpi.R")

---

Author

Mike Proctor – FlatBroke Farms