RT @MelissaLDell: I have a predoc opening! This position will involve working with historical U.S. local newspaper data to understand Ameri….

James Robinson (@HarrisPolicy) and I are hiring one or more predocs starting Summer '25. Come work with us on political economy, economic history, and AI. Link: @predoc_org.

RT @NobelPrize: BREAKING NEWS.The Royal Swedish Academy of Sciences has decided to award the 2024 Sveriges Riksbank Prize in Economic Scien….

RT @jjopperman: How A River Changing Course Led People To Form The First Governments. Research in @AEAjournals from Robert Allen, Mattia C.….

RT @MelissaLDell: We've released a new dataset, Newswire 2.7M unique newswires reproduced 32M+ times over a century….

🚨Job alert🚨@edoardoteso and I are looking for a predoc, for 1 or 2 years, to start immediately. Work with us on political economy, economic history, and machine learning. Apply here: @predoc_org.

RT @_alice_evans: What led to Iraq’s 9thC collapse?. Superb paper by Bob Allen & @LeanderHeldring. Rulers struggled for succession, & neede….

Very nice animation of our recent paper, thank you @econimate!.Paper: Economist article: Interview:

Once you're happy with your settings, we provide a routine to process a directory full of scans. Caveat: For your use case you will probably engineer our code a bit, but we hope it's a useful starting point.

Here is an example from a table in a historical data source. After cropping, our routine returns the tabular data in dataframe format, ready for processing.

Here's an example of a page from the historical New York Times that has text in various different boxes. We return the different text boxes. For a model that links different textboxes correctly, see my co-authored project:

An example of preprocessing is cropping to remove logos and other extraneous objects in an image. We also provide routines for:.1. Splitting of pages (e.g. multiple columns).2. Grayscale/contrast/brightness (for OCR contrast).These help a lot with the quality of the final OCR

We provide a notebook that allows for three tasks:.1. Digitization of images with (complex) text.2. Digitization of images with tables.3. Preprocessing of images.This utility is aimed at prototyping: Quickly trying out what you would need to do for large scale digitization.

New utility for prototyping document digitization. Website: Get started right away: 🧵: Tabular data, cropping, grayscale, done quickly.Developed with Michael Giordano (@KelloggSchool)

Empirical political economy’s impact on political philosophy. Nice full page article by @TheEconomist on our recent paper on the origins of government. Article: Paper: Interview:

RT @MelissaLDell: I’m excited to share News Déjà Vu (, which uses a custom large language model to retrieve histori….

RT @XJaravel: [1/4] Merci @gcalignon @natsilbert @LesEchos pour cette interview sur mon livre à paraître ce vendredi, qui analyze combien l….

We provide a lot of additional data at In archeology, we were particularly inspired by work by @jasonalikur Marc van de Mieroop, and Steven Garfinkle. Check their work. If you want to work with the data, reach out!.

We do not claim that all govt's are 'good'. We only make a claim about the origins of the first govt's. Historically, govt's have done terrible things. After gov is established, it needs some power to enforce. How to constrain that power is central to political economy research.

Nerdy caveat: Opportunities for coordination are common, but early states were rare. We do cost benefit analyses: Early state formation is a 'knife edge' result: Often, farmers just migrated but sometimes they creatively adapted existing social institutions into government.