Project Component 3 — Preliminary Analysis

LIN 301: Computation for Linguists

Author

Dr. Andrew M. Byrd

Overview

Now that you’ve collected and cleaned your data, it’s time to do some actual analysis. This component is preliminary, meaning you’ll get feedback before your final presentation.

Your job here is to use Python to analyze patterns or trends in your dataset using what you’ve learned so far — including pandas, WordNet, and spaCy (as appropriate to your data).

Goal

Use your cleaned .csv file from Component 2 to:

  • Load and explore your data;
  • Perform at least two types of quantitative or linguistic analysis;
  • Summarize and visualize your findings in a clear and interpretable way;
  • Explain what you found and what you think it means.

Deliverables

  • last_name_first_name_component3.qmd
  • last_name_first_name_component3.csv (if new columns or results are saved)
  • Optional: any figures you generate (e.g., in a /figures subfolder)

Instructions

1. Set Up

  • Work inside your existing course_project folder.
  • Load your cleaned CSV from Component 2 using pd.read_csv().
  • Confirm that your data appears as expected.

2. Plan Your Analysis

Before diving into the code, write a short paragraph (3–5 sentences) outlining:

  • What you want to learn from your data;

  • Which questions or hypotheses you hope to explore;

  • What results or trends you expect to find.

This short plan will help guide your coding decisions and make your later interpretation clearer.

3. Analyze Your Data

Perform at least two different analyses. Your exact methods will depend on your dataset, but here are some examples:

Type of Data Possible Analyses
Word lists Frequency counts, word length averages, vowel/consonant ratios
Text corpora POS distribution, named entity frequency, most common lemmas
Multilingual data WordNet synonym comparison, translation alignment, semantic similarity
Survey/numeric data Summary statistics, group comparisons, correlations

You might also:

  • Create a new column (e.g., word frequency, token length, POS tag, sentiment score);

  • Filter or group your data (e.g., by language, part of speech, or category);

  • Compute descriptive statistics with df.describe() or df.value_counts().

4. Visualize

Include at least one visualization using either:

  • matplotlib or pandas.plot(), or

  • any other tool you prefer (e.g., WordCloud, bar chart, histogram, etc.).

Be sure your plot is labeled and readable — axis labels, title, etc.

5. Interpret

Write a short paragraph (4–6 sentences) in your .qmd explaining:

  • What questions you asked of your data;

  • What methods or functions you used to answer them;

  • What your analysis revealed (summary of key results);

  • How you might extend or refine this analysis for your final project.

6. Save and Verify

  • If you created any new columns or computed results, save them as a new .csv.

  • Confirm your plots render properly in HTML.

  • Upload all required files to Canvas.

Grading Rubric (25 points total)

Category Description Points
Data Setup Properly loads and references Component 2 dataset; data verified and clean 5
Analysis Plan Clear and thoughtful description of goals, questions, and expected results 5
Analysis At least two clear, valid analyses with appropriate use of Python functions (pandas, spaCy, WordNet, etc.) 5
Visualization At least one clear, labeled, and interpretable chart or graphic 5
Interpretation 4–6 sentences explaining questions, methods, and findings; thoughtful reflection 5