On the Pomona server

1. Open https://rstudio.pomona.edu and log in
2. Go to File → New Project…
3. Choose New Directory → New Project
4. Enter a project name following the convention coursecode_topic — for example:

ea030_streams

5. For Create project as subdirectory of, leave the default (~) — this places the project in your home directory on the server
6. Click Create Project

On a local installation

Steps 2–6 are identical. The project folder appears wherever you chose in step 5.

One project per analysis

The EA030 standard structure

What goes where

Create the folders

Run this once in the Console after creating the project. It builds the entire skeleton instantly:

# Run once from the Console — creates all project folders
dirs <- c(
  "data/raw", "data/processed",
  "scripts",
  "output/figures", "output/tables", "output/reports"
)
invisible(lapply(dirs, dir.create, recursive = TRUE, showWarnings = FALSE))

# Verify
list.dirs(recursive = TRUE)

Or create them manually in the Files pane: click New Folder, name it, then navigate inside and repeat for subfolders.

The problem with absolute paths

An absolute path encodes the full location of a file from the root of the file system:

A relative path is measured from the working directory — and because .Rproj always sets the working directory to the project root, "data/raw/stream_sites.csv" resolves correctly whether you are on the Pomona server, your laptop, or a collaborator’s machine.

Relative paths like "data/raw/file.csv" work most of the time, but they depend on the working directory being set correctly. Inside a knitted .Rmd, the working directory is the folder containing the .Rmd file — not necessarily the project root. This causes subtle bugs when scripts are in a subfolder.

The here package always resolves paths from the project root, regardless of where the script lives. {#here}

library(here)

# here() returns the absolute path to the project root
here()
# "/home/jsmith/ea030_streams"

# Construct paths safely
read.csv(here("data", "raw", "stream_sites.csv"))

# Save output to a subfolder
ggsave(here("output", "figures", "do_vs_temp.png"),
       width = 7, height = 4)

Verify here is working

library(here)
here()   # should print the path ending in your project name

The three rules

1. Machine-readable — no spaces, no special characters, no accented letters
2. Human-readable — a glance tells you what the file contains
3. Sortable — numbers and ISO dates keep files in the right order

Numbering scripts

Number scripts with a two-digit prefix so they sort in run order in any file browser:

# Correct — sorts as 01, 02, 03 ... 10, 11
01_load_raw_data.R
02_clean_and_filter.R
03_exploratory_analysis.Rmd
04_linear_regression.R
05_report.Rmd

# Wrong — sorts as 1, 10, 11, 2, 3 (lexicographic order)
1_load.R
10_final.R
2_clean.R

Dates in filenames

When a file is versioned by date (raw data downloads, for example), use ISO 8601 format (YYYY-MM-DD) — it sorts chronologically:

# Good — sorts chronologically
stream_discharge_2024-03-15.csv
stream_discharge_2024-04-02.csv
stream_discharge_2024-11-01.csv

# Bad — ambiguous, sorts wrong
stream_discharge_3-15-24.csv
stream_discharge_Nov_data.csv

Script 01 — load and clean

Every analysis starts the same way: load raw data, inspect it, clean it, save a processed copy. No output figures yet — just getting the data ready.

# scripts/01_load_clean.R
# Purpose: load raw stream data, check, clean, and save processed copy
# Author:  your name
# Date:    2024-05-12

library(tidyverse)
library(here)
library(janitor)

# ── 1. Load ────────────────────────────────────────────────────────
raw <- read_csv(here("data", "raw", "stream_sites.csv"))

# ── 2. Inspect ─────────────────────────────────────────────────────
glimpse(raw)
summary(raw)

# ── 3. Clean ───────────────────────────────────────────────────────
clean <- raw %>%
  clean_names() %>%                       # snake_case column names
  filter(!is.na(dissolved_oxygen_mgl)) %>% # drop rows missing DO
  mutate(
    date       = as.Date(date, "%Y-%m-%d"),
    site_id    = factor(site_id),
    do_mgl     = dissolved_oxygen_mgl,
    temp_c     = water_temp_c
  ) %>%
  select(date, site_id, do_mgl, temp_c, flow_cms, conductivity_us)

# ── 4. Save processed copy ─────────────────────────────────────────
write_csv(clean, here("data", "processed", "stream_clean.csv"))
cat("Saved:", nrow(clean), "rows to data/processed/stream_clean.csv\n")

Script 02 — exploratory visualisation

# scripts/02_exploratory_plots.R
# Purpose: initial visualisations of cleaned stream data

library(tidyverse)
library(here)

clean <- read_csv(here("data", "processed", "stream_clean.csv"))

# ── DO vs temperature ──────────────────────────────────────────────
p_do_temp <- ggplot(clean, aes(temp_c, do_mgl, colour = site_id)) +
  geom_point(alpha = 0.55, size = 2) +
  geom_smooth(method = "lm", se = FALSE, linewidth = 0.8) +
  scale_colour_brewer(palette = "Dark2", name = "Site") +
  labs(x = "Water temperature (°C)", y = "Dissolved oxygen (mg/L)",
       title = "DO–Temperature Relationship by Site") +
  theme_minimal(base_size = 13) +
  theme(plot.title = element_text(family = "serif", face = "bold",
                                  colour = "#1e3a2f"))

# Save — never use Export button for reproducible figures
ggsave(here("output", "figures", "do_vs_temp.png"),
       plot = p_do_temp, width = 7, height = 4, dpi = 150)

cat("Figure saved to output/figures/do_vs_temp.png\n")

Script 03 — analysis report (Rmd)

---
title: "Stream Chemistry Analysis"
author: "Your Name"
date: "`r Sys.Date()`"
output:
  html_document:
    self_contained: true
    toc: true
---

```{r load, include=FALSE}
library(tidyverse)
library(here)
clean <- read_csv(here("data", "processed", "stream_clean.csv"))
```

```{r do-temp-plot, fig.cap="DO decreases with temperature across all sites.", echo=FALSE}
ggplot(clean, aes(temp_c, do_mgl, colour = site_id)) +
  geom_point(alpha = 0.55) +
  geom_smooth(method = "lm", se = FALSE) +
  theme_minimal()
```

Knit with Ctrl/⌘+Shift+K. The rendered HTML lands in the same folder as the .Rmd — move it to output/reports/ manually, or set the knit: option in the YAML to write there directly (as all EA030 tutorials do).

What’s next?

• Module 00 — R & RStudio Setup — server login and local installation
• Module 02 — R Markdown & knitr — YAML options, chunk options, and knitting to HTML
• Module 03 — Data Import — reading CSV, Excel, and field-data files