A full beginner curriculum for dry-lab biology projects: clean messy datasets, analyze them in RStudio or Python, and produce reproducible figures for an EBS poster.
By the end of this curriculum, students should be able to move from a biological question to a defensible, reproducible analysis.
Import CSV files, inspect variables, handle missing values, summarize groups, and choose appropriate statistical tests.
Work in RStudio for R-based statistics and in Python notebooks for pandas, NumPy, and SciPy workflows.
Keep an analysis log that explains cleaning decisions, assumptions, test selection, and limitations.
Create clear, labeled visualizations and export them with captions that state the result and interpretation.
Students can choose either track first. The website teaches both workflows directly; the only outside step is opening the tool you want to use for running code.
Use RStudio when your project is centered on statistical tests, group summaries, and clean publication-style plots.
Start RStudio TrackUse Python when your project needs programmable data cleaning, custom calculations, or notebook-based analysis.
Start Python TrackThe sequence below works for an EBS research team, club workshop, or independent student. Each module includes a concept, a hands-on lab, and a checkpoint deliverable.
Turn a biology interest into an analyzable question and identify whether the data are categorical, numeric, time-series, sequence-based, or text/literature-based.
Variables, observations, population, sample, outcome, predictor, and confounder.
Rewrite three broad biology topics as testable data questions.
Submit one research question plus the public dataset you plan to use.
Learn why analysis should happen in code instead of manual spreadsheet edits. Set up a folder structure for data, scripts, outputs, and notes.
Tidy data, raw versus cleaned files, relative paths, and analysis logs.
Import a CSV, inspect column names, preview rows, and save a cleaned copy.
Create a project folder with a README explaining the dataset source.
Identify missing values, duplicated rows, impossible values, inconsistent units, and mixed data types before running statistics.
Missingness, outliers, data dictionaries, unit standardization, and audit trails.
Clean a messy biological dataset and record every transformation.
Submit before/after row counts and a three-sentence cleaning rationale.
Summarize distributions with counts, proportions, mean, median, standard deviation, interquartile range, and grouped summaries.
Central tendency, spread, skew, sample size, and biological interpretation.
Build a grouped summary table comparing two biological conditions.
Write one paragraph explaining the most important summary result.
Choose figure types that fit the question: bar charts for counts, scatter plots for relationships, box plots for distributions, and line charts for time.
Axes, labels, units, color, uncertainty, figure captions, and avoiding misleading scales.
Create three figure drafts from the same dataset and choose the clearest one.
Export one poster-ready figure with a caption.
Select and run basic tests: chi-square, t-test, Mann-Whitney U, correlation, and simple linear regression.
Null hypothesis, p-values, effect size, confidence intervals, and assumptions.
Run two candidate tests and justify which one fits the data better.
Submit a test-selection note with assumptions and limitations.
Find data from public repositories and evaluate whether the metadata are complete enough for analysis.
Repository metadata, accession numbers, licensing, citation, and reproducibility.
Locate one dataset and write a short data provenance statement.
Add formal dataset citations to your analysis README.
Use simple models carefully: regression for relationships, classification for categories, and simulation for biological systems.
Training data, overfitting, validation, prediction versus explanation, and model limits.
Fit a simple regression model and compare predictions to observed values.
Explain whether your model supports a biological claim or only describes a pattern.
Connect results back to the biological question without overstating conclusions. Identify sampling bias, measurement limits, and alternative explanations.
Correlation versus causation, confounding, multiple comparisons, and scope of inference.
Write a limitations section for a published or sample analysis.
Submit three limitations and one next experiment or next analysis.
Organize the final project so another student can rerun the analysis and understand the result.
README files, commented code, output folders, citations, and figure exports.
Clean your project directory and rerun the notebook or script from start to finish.
Submit a reproducibility checklist with your abstract and poster draft.
The two tool tracks are now separate pages so each one can teach from first principles without crowding the course map.
Learn pandas for tables, NumPy for numeric arrays, and SciPy for statistical tests using an on-page biology dataset and guided exercises.
Open Python TrackLearn R data frames, formulas, grouped summaries, t-tests, and base plotting using the same built-in dataset and answer checks.
Open RStudio TrackThe curriculum ends with a small research package that can become the methods/results backbone of an EBS poster.
| Deliverable | What It Must Include | Quality Bar |
|---|---|---|
| Research question | One sentence naming the biological system, variables, population, and expected analysis. | Specific enough that another student could identify the needed dataset. |
| Clean dataset | Raw data source, cleaned file, and a note explaining exclusions or transformations. | No hidden manual edits. Cleaning steps are reproducible in code. |
| Analysis notebook or script | Readable RStudio script or Python notebook with imports, cleaning, summaries, tests, and plots. | Runs from top to bottom and produces the same result each time. |
| Figure and caption | One main result figure with labeled axes, units, legend if needed, and a caption. | The caption states the pattern and avoids overstating causation. |
| Interpretation paragraph | Result, biological meaning, limitation, and one next step. | Clear enough for a judge outside the exact subfield to follow. |
Use this checklist before moving from analysis into poster design.