Hey, I'm Kai!(he/they)

Statistician, Data Scientist & NLP researcher.

Introduction

I am a statistician and data scientist specializing in Natural Language Processing and AI, with a strong passion for using my knowledge to help us to better understand and quantify how societies communicate and evolve.

In my current research, I am focusing on analyzing political narratives - with a particular interest in how harmful discourse (including conspiracy theories) emerges, transforms, and spreads. My goal is to uncover patterns in the digital information flow and the forces shaping it by designing language models and applying them to corpora of news articles, political speeches or social media posts.

I aim to contribute research that helps to support a functional democracy. I am always interested in connecting with researchers, journalists, institutions, and organizations exploring the intersection of AI, politics, and society.

Kai-Robin Lange

A brief timeline

  • 2025–nowPostdoctoral researcher: detecting conspiracy theories in social media, spatio-temporal language models, and the German energy transition discourse. Also lecturing on NLP and text mining.
  • 2025–nowExternal Lecturer: teaching students of the Master Business Analytics at EBS University of Business and Law the fundamentals of Natural Language Processing and Text Mining.
  • 2021–2025Doctoral researcher: extracting economic narratives from text, temporal language modelling, NLP-based media analysis. Dissertation passed magna cum laude.
  • 2019–2021M.Sc. Statistics, thesis on unsupervised sentiment analysis, which later grew into Lex2Sent, my first published software package and paper.
  • 2016–2019B.Sc. Statistics at TU Dortmund University, with a minor in computer science.

What I work on

I design language models that make political and economic discourse measurable, then point them at news articles, parliamentary speeches and social media posts.

Conspiracy theories in social media

How conspiratorial narratives emerge, mutate and travel across platforms and what that says about the health of public debate.

Temporal & spatial language modelling

Treating text as a moving target: rolling topic models, diachronic embeddings and semantic shift detection over years of text.

Economic narratives

Extracting the stories economies tell about themselves, combining dynamic topic models with large language models.

The German energy transition

Tracking how the Energiewende is argued about in parliaments and the press, and how those arguments shift over time.

Recent papers, talks and distinctions.

Identifying economic narratives in large text corpora

talkInvited Talk

Mining a rabbit hole: leveraging language models to traverse conspiracy theory narratives in telegram channels

talkInvited Talk

Identifying economic narratives in large text corpora

talk

Do Investors' Mental Models Transfer Across Markets?

joint with Paul Grass

grant

Following Socio-Environmental Conflict Narratives About Energy Transition in Chile: A Spatio-Temporal Analysis Using Dynamic Topic Modeling

publicationPDF

Word-Centered Semantic Graphs for Interpretable Diachronic Sense Tracking

publicationDOI

Open Source Software

GitHub

Libraries and corpora anyone can use.

ttta: Tools for Temporal Text Analysis

Python Library · released 09/2024

GitHubPyPi

Lex2Sent

Python Library · released 03/2023

GitHubPyPi

Get in touch

The quickest way to reach me is by email. I am always glad to hear from researchers, journalists, institutions, companies, and organizations working at the intersection of AI, politics and society about a joint project, a guest lecture or workshop, a data-driven story, or a thesis topic.

website@krlange.de GitHub LinkedIn Google Scholar