---
title: README
emoji: ⚡
colorFrom: gray
colorTo: gray
sdk: static
pinned: false
---

# Jerzak Labs

_Data, models, and methods for planetary-scale causal inference, text-based AI systems, and other interests._
Led by [**Connor T. Jerzak**](https://connorjerzak.com), Assistant Professor at UT Austin.

<a href="https://planetarycausalinference.org/book-launch">
  <picture>
    <source media="(prefers-color-scheme: dark)" srcset="https://planetarycausalinference.org/wp-content/uploads/2025/11/PCI_book_dark.png">
    <img src="https://planetarycausalinference.org/wp-content/uploads/2025/11/PCI_book.png" alt="Book" width="200">
  </picture>
</a>


---

## Recent Team Tutorials

-   [**Nicolas Audinet de Pieuchon**](https://nicaudinet.github.io/) presents: [*Benchmarking Debiasing Methods for LLM-based Parameter Estimates*](https://connorjerzak.com/transcript-benchmarking-debiasing-methods/)
-   [**Nicolas Audinet de Pieuchon**](https://nicaudinet.github.io/) presents: [*Can Large Language Models (or Humans) Disentangle Text?*](https://connorjerzak.com/transcript-disentangle-video/)
-   [**Adel Daoud** ](https://adeldaoud.com/)presents: [*A First Course in Planetary Causal Inference: Confounding*](https://planetarycausalinference.org/transcript-pci-tutorial-2025/) (@IC2S2 2025)
-   [**Adel Daoud** ](https://adeldaoud.com/)presents: [*Planetary Causal Inference*: Overview](https://planetarycausalinference.org/transcript-pci-seminar-yale/) (@Yale)
-   [**Connor Jerzak**](https://connorjerzak.com/) presents: [*Seeing Like a Satellite While Learning Across Scales: Remote Audits + Multi-Scale Optimization for Heterogeneity*](https://planetarycausalinference.org/transcript-seeing-like-a-satellite/) (@Columbia)
-   [**Connor Jerzak**](https://connorjerzak.com/) presents: *[Selecting Optimal Candidate Profiles in Adversarial Environments](https://connorjerzak.com/transcript-adversarial-2024/)* (@UT Dallas & National Chung Hsing University)
-   [**Richard Johansson**](https://www.cse.chalmers.se/~richajo/) presents: *[Conceptualizing Treatment Leakage in Text-based Causal Inference](https://connorjerzak.com/transcript-leakage-video/)* (@NAACL)
-   [**Satiyabooshan Murugaboopathy**](https://www.linkedin.com/in/msatiya/?originalSubdomain=de) presents: *[Platonic Representations for Poverty Mapping: Unified Vision-Language Codes or Agent-Induced Novelty?](https://aidevlab.org/transcript-platonic/)*
-   [**Kazuki Sakamoto**](https://www.linkedin.com/in/kazukisakamoto/) presents: *[A Scoping Review of Earth Observation and Machine Learning for Causal Inference](https://connorjerzak.com/transcript-a-scoping-review-pci/)*
-   [**Fucheng Warren Zhu**](https://www.warrenzhu.com/) presents: [*Optimizing Multi-Scale Representations to Detect Effect Heterogeneity Using EO and Computer Vision*](https://connorjerzak.com/transcript-encoding-multi-level-dynamics-in-effect-heterogeneity-estimation/)

---

## Research Areas

- **[Planetary Causal Inference](https://planetarycausalinference.org/)**  
  Harnessing Earth observation data, randomized control trials, and advanced modeling for global development and climate impact.

- **[Research Design](https://connorjerzak.com/research-design/)**  
  Rerandomization procedures, effect heterogeneity, and leveraging images (e.g., satellite) as data sources.

- **[Text-based AI Systems](https://connorjerzak.com/text-overview/)**  
  Automated nonparametric text analysis, large language models, and language-based causal inferences.

- **[Descriptive Representation & Political Economy](https://globalleadershipproject.net/)**  
  Understanding group-level representation worldwide; analyzing policy impacts.

---

## Featured Repositories

- **[CausalImages](https://github.com/cjerzak/causalimages-software)**  
  R Package for performing causal inference with images, including Earth observation and biomedical data.

- **[DescriptiveRepresentationCalculator](https://github.com/cjerzak/DescriptiveRepresentationCalculator-software)**  
  Tools for measuring descriptive representation across gender, ethnicity, religion, and more in global leadership data.

- **[readme2](https://github.com/iqss-research/readme-software)**  
  Enhanced automated content analysis for social science text data (with [paper](https://gking.harvard.edu/files/gking/files/word.pdf)).

- **[LinkOrgs](https://github.com/cjerzak/LinkOrgs-software)**  
  Linking organizational datasets using half-a-billion open-collaborated records.  

For additional code packages, see [Code Overview](https://connorjerzak.com/software/).

---

## Data Assets

Explore large-scale data on Earth observation + RCTs, text-based sentiment disentanglement, organizational name-matching, and more:

- **[HumanDisentangledText](https://huggingface.co/datasets/cjerzak/HumanDisentangledText)**  
- **[LinkOrgs](https://huggingface.co/datasets/cjerzak/LinkOrgs)**  
- **[ImageHeterogeneity](https://huggingface.co/datasets/cjerzak/ImageHeterogeneity)**  

Further details: [Data Overview](https://connorjerzak.com/data/).

---

## Contact

- Email: [connor.jerzak@austin.utexas.edu](mailto:connor.jerzak@austin.utexas.edu)  
- Website: [ConnorJerzak.com](https://connorjerzak.com)  
- GitHub: [GitHub.com/cjerzak](https://github.com/cjerzak)
- Subscribe: [Jerzak Labs YouTube](https://www.youtube.com/@connorjerzak?sub_confirmation=1)
- Subscribe: [Planetary Causal Inference](https://www.youtube.com/@PlanetaryCausalInference?sub_confirmation=1)