Planorama

A Good Plan is Hard to Find: Aligning Models with Preferences is Misaligned with What Helps Users
EMNLP 2025

This repository contains the code for Planorama, the interface developed for our EMNLP 2025 paper.

📄 Our paper: https://nbalepur.github.io/assets/pdf/Planorama.pdf
📊 Our dataset of preferences and outcomes when using plans in Planorama: https://huggingface.co/datasets/nbalepur/Planorama-user-data)

---

🚀 Local Development

1. PostgreSQL Setup

Install PostgreSQL. A quick option is via Conda:

conda install anaconda::postgresql

Initialize a database cluster (⚠️ WSL users: prefer a subdirectory under $HOME):

mkdir datadir
pg_ctl -D datadir initdb
pg_ctl -D datadir -l logfile start

Create the user and database:

createuser -s postgres
createdb -U postgres -h localhost -p 5432 kuiperbowl

Then, configure web/.env.local from .env with your credentials.

---

2. Application Setup

Create and activate a virtual environment:

conda create -n helpfulness
conda activate helpfulness
conda install pip

Install dependencies & set up:

cd web
pip install -r requirements.txt
python manage.py migrate

Start Redis for channel layers (via Docker):

docker run -p 6379:6379 -d redis:5

Load fixtures:

python manage.py loaddata fixtures/question_fixtures.json
python manage.py loaddata fixtures/document_fixtures.json

Run the server:

python manage.py runserver --insecure

---

📚 Entering Question Data

Our math and trivia questions can be loaded directly from the repository:

cd web
python scripts/pb_load.py
python manage.py loaddata fixtures/sanity_tutorial_questions.json
python manage.py loaddata fixtures/question_fixtures.json

---

🐳 Using Docker

Quick Start

1. Configure .env with the correct credentials.
2. Build and run:

docker-compose up --build

---

📖 Citation

If you use this code or dataset, please cite our EMNLP 2025 paper:

@inproceedings{balepur2025planorama,
  title={A Good Plan is Hard to Find: Aligning Models with Preferences is Misaligned with What Helps Users},
  author={Balepur, N. and others},
  booktitle={Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP)},
  year={2025}
}