Mini Project 3: Beyond the Joke

INFO 230 | Spring 2026

How do US political and COVID memes differ in how they assign blame, victimhood, and heroism. Do visual, textual, and acoustic features reinforce those framings? This project applies a full multimodal cultural analytics pipeline across three datasets and eight steps of analysis to find out.

For the full analysis, results, discussion, and interpretations see code/project-workflow.ipynb.

cultural-analytics-project-3/
├── code/
│   ├── project-workflow.ipynb       <- full analysis lives here
│   └── environment.yaml             <- conda environment
├── data/
│   ├── audio_data/                  <- M-Arg debate audio dataset
│   │   ├── files_explanation.md
│   │   ├── M-arg_preprocessing.py
│   │   └── data/
│   │       ├── split audio/         <- 17 MP3 debate clips
│   │       ├── timestamps/          <- force-aligned utterance timestamps
│   │       └── tokenised text/      <- speaker-labeled transcripts
│   ├── data_clean/                  <- 2020 campaign speech TSVs
│   ├── SampleImagesData/
│   │   └── SampleImagesData/
│   │       ├── covidMemes/          <- 8 COVID meme PNGs
│   │       └── usPoliticsMemes/     <- 8 US Politics meme PNGs
│   ├── text_with_ocr.csv            <- meme training set (5,552 rows)
│   └── validation_data.csv          <- meme validation set (650 rows)
├── README.md
└── index.html

1. Clone the repo
2. Create and activate the conda environment:

conda env create -f environment.yaml
conda activate cultural_analytics

3. Run code/project-workflow.ipynb top to bottom

All dependencies are managed via environment.yaml using the cultural_analytics conda environment. Key packages include: pandas, numpy, nltk, scikit-learn, plotly, matplotlib, Pillow, opencv-python, librosa

yogesh239. (2023). Memes Images: OCR Data. Kaggle. https://www.kaggle.com/datasets/yogesh239/text-data-ocr

Chalkiadakis, I., Angles d'Auriac, L., Peters, G.W., and Frau-Meigs, D. (2025). A text dataset of campaign speeches of the main tickets in the 2020 US presidential election. Scientific Data, 12, 662. https://doi.org/10.1038/s41597-025-04681-x | Dataset: https://doi.org/10.5281/zenodo.14785782

Mestre, R., Milicin, R., Middleton, S.E., Ryan, M., Zhu, J., and Norman, T.J. (2021). M-Arg: Multimodal Argument Mining Dataset for Political Debates with Audio and Transcripts. Proceedings of the 8th Workshop on Argument Mining, pp. 78-88. Association for Computational Linguistics. https://aclanthology.org/2021.argmining-1.8/ | Dataset: https://zenodo.org/records/5653504

Chowdhury, J.H., Ramanna, S., and Kotecha, K. (2025). Speech emotion recognition with light weight deep neural ensemble model using hand crafted features. Scientific Reports, 15, 11824. https://doi.org/10.1038/s41598-025-95734-z

Hutto, C.J. and Gilbert, E.E. (2014). VADER: A Parsimonious Rule-Based Model for Sentiment Analysis of Social Media Text. Proceedings of the 8th International AAAI Conference on Weblogs and Social Media (ICWSM-14). https://doi.org/10.1609/icwsm.v8i1.14550

McFee, B., Raffel, C., Liang, D., Ellis, D.P.W., McVicar, M., Battenberg, E., and Nieto, O. (2015). librosa: Audio and Music Signal Analysis in Python. Proceedings of the 14th Python in Science Conference (SciPy 2015), pp. 18-24. https://doi.org/10.25080/majora-7b98e3ed-003

Claude (Anthropic) was used as an AI assistant throughout this project for help with code structure, debugging, implementation decisions, and writing feedback. All code was read through, understood, and retyped manually. All analytical decisions, interpretations, and conclusions are my own. Claude helped, but the thinking was mine.

By Maddie MacDonald | UC Berkeley Statistics, MA