This wiki retires in 2027; content deletion started in 2026. No planned cloud migration.
|
-
Created by
InKyung Choi, last updated by Stela Derivolcov on 17 Jul, 2024
2 minute read
This page contains generative AI use cases in national and international statistical organizations presented at various relevant meetings. If you want your use case / research to be added, please submit it via this form
| Use case type | Target user | Title | Resource type | Organisation | Date | Models | Approach | Tools | Code availability |
|---|---|---|---|---|---|---|---|---|---|
| Code translation / generation | Internal | Report | Ireland CSO | 2023-12 |
|
|
|
| |
| Code translation / generation | Internal | SAS to R conversion: the use of LLM | Presentation | France Insee | 2023-11 |
|
|
| |
| Text generation - analysis report | Report | Canada | 2023-11 |
|
|
|
| ||
| Text generation - analysis report | External | Presentation | Statistics Norway | 2024-05 | GPT 4 | Prompt engineering | Chat GPT |
| |
| Text edit | Internal | Leveraging AI-assistants for better metadata and communication | Presentation | BIS | 2024-03 | GPT 3.5 | Prompt engineering, RAG | OpenAI AI Assistant |
|
| Communication materials - texts | Internal / External | Use of AI in the communication and dissemination of statistics | Indonesia BIS | 2024-03 | GPT 3.5 | Prompt engineering | ChatGPT | N/A | |
| Communication materials - images | Internal / External | Use of Generative AI in ESTAT | Presentation | Eurostat | 2024-04 | Prompt engineering | DALL.E, ChatGPT 4, Adobe Firefly, Image Creator, Midjourney | N/A | |
| Information search | UNCTAD |
| RAG |
|
| ||||
| Information search | External | Using large language models (LLMs) to improve website search experience with StatsChat | Report | UK ONS | 2023-08 | Flan-T5; all-mpnet-base-v2 | RAG | Flask | |
| Information search | External | OECD | GPT 3.5 | RAG | MILA wrapper? |
| |||
| Data search | External | IMF |
|
|
|
| |||
| Communication materials - texts, Information search, Summarize function | Internal | Presentation | Bundesbank Germany | 2024-06 | GPT 3.5, GPT 4, GPT 4 Turbo | Prompt engineering | Chat GPT via Azure Microsoft Cloud Service |
| |
| Communication material - video, audio | External | video | Statistics Canada | 2024-07 | Video/audio generation | HeyGen, ElevenLabs, ChatGPT, Hedra, Suno |
|
Typology for use case type
| Use case type | Examples |
|---|---|
| Code translation / generation | code translator from SAS to R |
| Synthetic data generation | textual description of statistical classification system codes |
| Communication materials - texts | press release, social media post |
Communication materials - images or videos | |
Drafting analysis report | |
| Text analysis for micro-level data processing | classification of text description into statistical classification system codes |
Text analysis for research purpose | market analysis from news papers |
Data search | converting data search through SDMX and LLM |
Information search | content search from organization’s website, internal documents |
Assisting in administrative tasks | internal email, meeting notes, preparing presentation slides |