HW 7 — Alice Across Languages

Final Group Project

Author

Dr. Andrew M. Byrd

Overview

This is your final group project and final homework for LIN 301: Computation for Linguists.
You will apply everything you’ve learned in this class — text cleaning, tokenization, frequency analysis, syntax, semantics, visualization, and collaboration — to analyze Alice’s Adventures in Wonderland across multiple languages.

Group Setup and GitHub Repository

You’ll be working in groups with a maximum of 4 students.

Group Size	Languages Analyzed
1 student	2 languages total — English + any one: French, Spanish, Finnish, Italian
2 students	3 languages total — English + any two: French, Spanish, Finnish, Italian
3 students	4 languages total — English + any three: French, Spanish, Finnish, Italian
4 students	5 languages total — English, French, Spanish, Finnish, Italian

Each group must create a shared GitHub repository for the project.
Your repo should be named something like: alice-multilingual-group

Repository Requirements

Create the repo
- One member creates the repo on GitHub.
- Add all teammates as collaborators (Settings → Collaborators).
- Clone it locally through RStudio:
  File → New Project → Version Control → Git.
Folder structure

Repository structure

alice-multilingual-group/
├── data/             # text files (raw and cleaned)
├── figures/          # visualizations
├── HW7_alice.qmd # main Quarto notebook
├── README.md         # project summary and setup notes
└── requirements.txt  # Python packages needed to run the notebook

README contents

Group member names and roles
Languages analyzed
Instructions to reproduce results
Dependencies or model names (en_core_web_sm, fr_core_news_sm, etc.)

Collaboration

Everyone commits code regularly (not just one person).
Commit messages should describe what changed.
In fairness of those students who do not use LLMs, please avoid the use of these as a tool. You may copy code that we’ve used previously in the class or write your own.

Downloading Files

You may download the following files from the below websites (using curl or python code).

English - https://www.gutenberg.org/cache/epub/11/pg11.txt
French - https://www.gutenberg.org/cache/epub/55456/pg55456.txt
Finnish - https://www.gutenberg.org/cache/epub/46569/pg46569.txt
Italian - https://www.gutenberg.org/cache/epub/28371/pg28371.txt
Spanish - https://github.com/andergd/separadorSilabas/blob/master/Carroll%2C%20Lewis%20-%20Alicia%20En%20El%20Pa%C3%ADs%20De%20Las%20Maravillas.txt

Core Components

Each group’s notebook must include four major sections. You should assign roles to each member of your group for maximal fairness.

1. Text Cleaning & Setup

Manually remove headers and footers. Then, load and clean all language files and tokenize using the appropriate spaCy pipeline for each language.

Compute and compare basic corpus stats:

total tokens
unique lemmas
lexical diversity (type-token ratio), which measures how varied the vocabulary is
number of sentences

Summarize these metrics in a small comparative table.

2. Frequency & Collocations

Lemma Frequency:

Top 20 content lemmas per language (exclude stopwords and punctuation).
Visualize results in a word cloud for each language. Label each visualization clearly.

Discuss any major differences you see between languages, if you find any.

(Optional) You may also include a single bar chart comparing top lemmas across languages for a more quantitative view.

3. Syntax Patterns

Use spaCy’s dependency parser to explore syntax.

Tasks:

Extract verbs where the name Alice is the subject (nsubj) or object (dobj).
- Count and rank the most frequent verbs across languages, sorting the values with the most frequent first.
- Create a table summarizing the top five verbs in each language.
Find adjectives modifying the name Alice.
- Count and rank the most frequent adjectives across languages, sorting the values with the most frequent first.
- Create a table summarizing the top adjectives.

Discuss what your findings perhaps reveals about the language in question or the translation.

4. Semantics & WordNet

Dive into meaning and concept structure.

Tasks:

Choose 10 shared lemmas (5 nouns, 5 verbs).
For each language, retrieve WordNet synsets, glosses, and hypernyms.
For one pair of nouns and one pair of verbs, compute similarity scores (e.g., Wu–Palmer) within each language.
Visualize semantic similarity within two scatter plots (one for nouns, one for verbs), which will show how similar or distinct the chosen word pairs are within and across languages.

Note: if WordNet coverage is limited for one of your chosen languages, describe what you attempted and how you adapted (e.g., using English synsets for comparison).

Note: Visualization & Interpretation

Present your findings clearly and creatively.

Include figure captions and axis labels.
Make sure all visuals are generated in the qmd (no pasted images).
Each figure should directly support a linguistic observation.

Include at least one meaningful visualization per major section.

Deliverables

Your GitHub repo must contain:

HW7_alice.qmd — main notebook (runs end-to-end)
README.md — repo overview and group info
/data/ and /figures/ folders with organized files

All text files must be saved as UTF-8 encoded .txt.
All visuals should be exported as .png or .svg into /figures/.

Regular commits by all members

Grading Rubric

Category	Weight	Description
Original, working code	25%	Student-written code that runs cleanly from a clone of the repo.
Repository organization & version control	20%	Clear folder structure, README, regular commits from all members.
Data processing & analysis	25%	Proper text cleaning, tokenization, syntax, and semantic work.
Visualizations & interpretation	20%	Clear, interpretable visuals and thoughtful linguistic discussion.
Collaboration & communication	10%	Evidence of shared work; consistent tone and structure across sections.

Submission

Make sure your repo is public and all code executes without errors.
- Submit the GitHub repo URL on Canvas by the due date.
Only one group member needs to submit the link, but all names must appear in the README.
Late submissions will lose points for collaboration and reproducibility.

--- title: "HW 7 — *Alice* Across Languages" subtitle: "Final Group Project" author: "Dr. Andrew M. Byrd" format: html: toc: true toc-depth: 2 code-tools: true fontsize: 12pt execute: freeze: auto warning: false message: false editor: markdown: wrap: 72 --- # Overview - This is your **final group project** and **final homework** for *LIN 301: Computation for Linguists*.\ - You will apply everything you’ve learned in this class — text cleaning, tokenization, frequency analysis, syntax, semantics, visualization, and collaboration — to analyze *Alice’s Adventures in Wonderland* across **multiple languages**. ------------------------------------------------------------------------ # Group Setup and GitHub Repository You’ll be working in groups with a maximum of **4** students. | Group Size | Languages Analyzed | |-----------------|------------------------------------------------------| | 1 student | 2 languages total — English + any one: French, Spanish, Finnish, Italian | | 2 students | 3 languages total — English + any two: French, Spanish, Finnish, Italian | | 3 students | 4 languages total — English + any three: French, Spanish, Finnish, Italian | | 4 students | 5 languages total — English, French, Spanish, Finnish, Italian | Each group must create a **shared GitHub repository** for the project.\ Your repo should be named something like: `alice-multilingual-group` ### Repository Requirements 1. **Create the repo** - One member creates the repo on GitHub. - Add all teammates as collaborators (Settings → Collaborators). - Clone it locally through RStudio:\ `File → New Project → Version Control → Git`. 2. **Folder structure** ::: callout-note #### Repository structure ```text alice-multilingual-group/ ├── data/ # text files (raw and cleaned) ├── figures/ # visualizations ├── HW7_alice.qmd # main Quarto notebook ├── README.md # project summary and setup notes └── requirements.txt # Python packages needed to run the notebook ``` ::: 3. **README contents** - Group member names and roles - Languages analyzed - Instructions to reproduce results - Dependencies or model names (`en_core_web_sm`, `fr_core_news_sm`, etc.) 4. **Collaboration** - Everyone commits code regularly (not just one person). - Commit messages should describe what changed. - In fairness of those students who do not use LLMs, please avoid the use of these as a tool. You may copy code that we've used previously in the class or write your own. ------------------------------------------------------------------------ # Downloading Files You may download the following files from the below websites (using `curl` or python code). - **English** - https://www.gutenberg.org/cache/epub/11/pg11.txt - **French** - https://www.gutenberg.org/cache/epub/55456/pg55456.txt - **Finnish** - https://www.gutenberg.org/cache/epub/46569/pg46569.txt - **Italian** - https://www.gutenberg.org/cache/epub/28371/pg28371.txt - **Spanish** - https://github.com/andergd/separadorSilabas/blob/master/Carroll%2C%20Lewis%20-%20Alicia%20En%20El%20Pa%C3%ADs%20De%20Las%20Maravillas.txt ------------------------------------------------------------------------ # Core Components Each group’s notebook must include **four major sections**. You should assign roles to each member of your group for maximal fairness. ------------------------------------------------------------------------ ## 1. Text Cleaning & Setup Manually remove headers and footers. Then, load and clean all language files and tokenize using the appropriate spaCy pipeline for each language. Compute and compare basic corpus stats: - total tokens - unique lemmas - lexical diversity (type-token ratio), which measures how *varied* the vocabulary is - number of sentences Summarize these metrics in a small comparative table. ------------------------------------------------------------------------ ## 2. Frequency & Collocations Lemma Frequency: - Top 20 content lemmas per language (exclude stopwords and punctuation). - Visualize results in a word cloud for each language. Label each visualization clearly. Discuss any major differences you see between languages, if you find any. (Optional) You may also include a single bar chart comparing top lemmas across languages for a more quantitative view. ------------------------------------------------------------------------ ## 3. Syntax Patterns Use spaCy’s dependency parser to explore syntax. Tasks: - Extract verbs where the name *Alice* is the **subject** (`nsubj`) or **object** (`dobj`).\ - Count and rank the most frequent verbs across languages, sorting the values with the most frequent first.\ - Create a table summarizing the top five verbs in each language.\ - Find adjectives modifying the name *Alice*.\ - Count and rank the most frequent adjectives across languages, sorting the values with the most frequent first.\ - Create a table summarizing the top adjectives. Discuss what your findings perhaps reveals about the language in question or the translation. ------------------------------------------------------------------------ ## 4. Semantics & WordNet Dive into meaning and concept structure. Tasks: - Choose **10 shared lemmas** (5 nouns, 5 verbs). - For each language, retrieve WordNet synsets, glosses, and hypernyms. - For **one** pair of nouns and **one** pair of verbs, compute similarity scores (e.g., Wu–Palmer) within each language. - Visualize semantic similarity within two scatter plots (one for nouns, one for verbs), which will show how similar or distinct the chosen word pairs are within and across languages. Note: if WordNet coverage is limited for one of your chosen languages, describe what you attempted and how you adapted (e.g., using English synsets for comparison). ------------------------------------------------------------------------ ## Note: Visualization & Interpretation Present your findings clearly and creatively. - Include figure captions and axis labels. - Make sure all visuals are generated in the qmd (no pasted images). - Each figure should directly support a linguistic observation. Include at least one meaningful visualization per major section.\ At the end, your final section should interpret the results linguistically: - What do the frequency and syntax patterns suggest?\ - Are some translations more literal, others more interpretive?\ - How do semantic networks differ across languages? ------------------------------------------------------------------------ # Deliverables Your **GitHub repo** must contain: 1. `HW7_alice.qmd` — main notebook (runs end-to-end) 2. `README.md` — repo overview and group info 3. `/data/` and `/figures/` folders with organized files - All text files must be saved as UTF-8 encoded `.txt`. - All visuals should be exported as `.png` or `.svg` into `/figures/`. 4. Regular commits by all members ------------------------------------------------------------------------ # Grading Rubric | Category | Weight | Description | |-------------------------|-------------------|----------------------------| | **Original, working code** | 25% | Student-written code that runs cleanly from a clone of the repo. | | **Repository organization & version control** | 20% | Clear folder structure, README, regular commits from all members. | | **Data processing & analysis** | 25% | Proper text cleaning, tokenization, syntax, and semantic work. | | **Visualizations & interpretation** | 20% | Clear, interpretable visuals and thoughtful linguistic discussion. | | **Collaboration & communication** | 10% | Evidence of shared work; consistent tone and structure across sections. | ------------------------------------------------------------------------ # Submission - Make sure your repo is public and all code executes without errors. - Submit the **GitHub repo URL** on Canvas by the due date. - Only one group member needs to submit the link, but all names must appear in the README. - Late submissions will lose points for collaboration and reproducibility. ------------------------------------------------------------------------