agronomic-data-automation-r

Automated data processing pipeline for multi-trait plant breeding trials. Uses functional programming in R to summarize plot-level statistics (Mean, SD, N) and analyze inter-plot variability

Automated Agronomic Data Pipeline

This project provides a robust framework for automating the summarization of field trial data. It transforms raw "Long Format" field observations into a structured "Wide Format" summary report, suitable for variety selection and breeding analysis.

Project Overview

In agricultural research, data is often collected per plant but must be analyzed per Plot. "Project Connor" automates this aggregation for multiple traits including:

• Yield Components: Capsule Count, Caps Count Per Plant.
• Morphology: Plant Height, Internode Length, First Capsule Height.
• Management: Harvest Date, Plant Count at Harvest.

Engineering Highlights

• Functional Automation: Developed the summarise_trait() function to eliminate repetitive code, using dynamic column naming (!!paste0).
• Data Integration: Utilizes full_join chains to merge disparate trait summaries into a single master ledger.
• Variability Analysis: Calculates Within-group vs. Between-group variability to assess the uniformity of experimental plots.
• Statistical Comparison: Integrated t-tests to compare specific plots against the trial average.

Visual Analytics

The pipeline generates diagnostic plots to visualize trait distribution across plots, identifying outliers and performance gaps.

Tech Stack

• Language: R
• Libraries: tidyverse (dplyr, ggplot2, readr)
• Methods: Functional Programming, Descriptive Statistics, Variance Analysis.

Repository Structure

• connor_pipeline.R: The core automation script.
• data/raw_field_data.csv: Sample input file (plot, trait, value).
• final_summary_traits.csv: The generated master report.