Syllabus
Course Learning Goals
- Give graduate students in agriculture, ecology, and related sciences an overview of multivariate datasets and their analyses.
- Develop proficiency in coding and conducting statistical analyses in R.
- Prepare students for further individual learning according to their specific needs.
- Promote collaborative discovery and resourcefulness.
Topics Covered
See the Schedule for week-by-week details.
| Week | Topic | Key Packages |
|---|---|---|
| 1 | R intro: objects, vectors, data frames, basic ggplot2 | ggplot2 |
| 2 | Linear regression: lm(), residuals, predictions, R², F-statistics, interactions |
ggplot2, plotly |
| 3 | General/Generalized linear models (GLM) | base R |
| 4 | Nonlinearity, GAMs, writing functions | mgcv |
| 5 | Model selection: ANOVA-based forward selection, AIC/BIC, mixed effects | MuMIn, MASS |
| 6 | Regularization (Ridge, Lasso, Elastic Net), cross-validation, AUC | glmnet |
| 7 | PCA and dimensionality reduction | reshape2, data.table |
| 8 | Clustering (K-means, hierarchical), apply functions | base R |
| 9 | MANOVA, random forests, neural nets, transformations | caret, randomForest, neuralnet |
| 10 | Structural equation modeling (SEM), redundancy analysis (RDA) | lavaan, piecewiseSEM, vegan |
Course Organization
- Lectures (Mon/Wed 1:10–2:30 PM): Background presentation + live coding tutorial. Bring your laptop.
- Discussion (Fri 12:10–1:00 PM): Problem set Q&A, bonus material, catch-up.
- Zoom: Available for all sessions. Recordings posted within 24hrs (best-effort; attend in person when possible).
Assignments & Grading
10 weekly homework sets + 1 final assignment. Submitted through Canvas.
Grading scale: Standard UC Davis A–F scale.
Textbooks & Resources
No required textbook. Recommended free resources:
Policies
- Attendance: In-person attendance strongly encouraged; Zoom available but not guaranteed.
- Reproducibility: All submitted R code must use
set.seed()before any random operations. - Academic integrity: Collaboration encouraged on problem sets; submitted work must be your own.