Hub How-Tos

How prevalent are AI bots in online samples, and can you spot human subjects using large language models (LLMs) to complete surveys? In this session, Nick Stagnaro, a survey researcher at Stanford and MIT, will discuss his recent research on the current state of AI bots in online samples.

Data providers understandably want to limit your use of AI with their data. We are committed to licensing data so that it can be used with AI. Learn about the limits and horizons for AI use in social science research.

Large language models (LLMs) are revolutionizing how we convert raw documents into structured datasets—but how well do they handle messy, real-world data? Join the Data, Analytics, and Research Computing (DARC) team to explore a benchmark dataset for information extraction.

Stanford researchers can now access SNL Insurance Data from S&P Global Market Intelligence. In this session, we introduce the data and review how it is structured and accessible in Redivis and Capital IQ Pro.

Topics include: SNL insurance data, LLMs for information extraction, using AI with licensed resources, and mitigating AI threats in survey research. Register to attend.

Should you use ChatGPT to debug errors, or let an AI agent write your code from scratch? Join the DARC team to navigate the evolving landscape of AI coding tools, discover Stanford’s best practices, and learn how to keep your licensed data secure.

This presentation compares and contrasts three datasets that allow for historical research into public and private firms both in the U.S. and around the world, including at the establishment level.

Enhance your Qualtrics surveys with an embedded AI chat window that participants can use directly within your survey. Learn about the secure, easy-to-implement tool developed by the Behavioral Lab and DARC teams.

Topics include: AI integration in Qualtrics, comparisons of historical private company and establishment-level datasets, AI code development, and newly added Verisk/Infutor data spanning five additional years. Register to attend.

A step-by-step guide to fine-tuning an open source model in the cloud and running inference on Stanford’s on-prem research computing platforms. Watch recording.

CloudResearch’s new tool streamlines running longitudinal studies on their platform. Learn how to set it up and see it in action. Watch recording.

Learn about how Stanford’s flagship clusters, Sherlock and Marlowe, can help meet your GPU-resource needs. Watch recording.

Dewey Data Founder & CEO, Evan Barry, introduces the platform and highlights datasets, such as LinkedIn and other workforce analytics; consumer data; and more. Watch recording.

Topics include: Dewey Data, GPU resources across Stanford, longitudinal surveys with CloudResearch, and fine-tuning LLMs. Watch recordings.

Stanford faculty have the opportunity to use their research funds for the Equifax Ignite Platform to gain access to individual-level consumer financial behaviors. Watch recording.

Explore common patterns in network activity and develop a clearer understanding of the potential and limitations of using present and historical web usage data. Watch recording.

Learn how to deploy LLMs on Stanford’s computing platforms. We also share practical tips for efficient inference on Ollama. Watch recording.

Stanford now has access to MSCI Climate data. Learn about this new dataset and how to work with it on the Redivis platform with example queries and workflows. Watch recording.

Topics include: MSCI Climate data, large language models on-prem, understanding network traffic data, and the Equifax Ignite platform. Watch recordings.