Data Science Dashboards

View slides in full screen

Dashboards are complex

• Many possible user inputs
• You don’t want the LLM to completely generate the application

Querychat

Demo: Querychat basics

https://shiny.posit.co/py/templates/querychat/

Demo: Querychat dashboard

https://shiny.posit.co/py/templates/sidebot/

You can get started

Use a free local llama model

• Free, local
• Models are not as good out of the box compared to other providers

Python Chatlas

import chatlas as clt
chat = clt.ChatOllama(model="qwen3:0.6b")
chat.chat("what is the capital of the moon?")

R Ellmer

library(ellmer)
chat <-  chat_ollama(model="qwen3:0.6b")
chat$chat("what is the capital of the moon?")

GitHub Model

• You will need to create a GitHub Personal Access Token (PAT).
• It does not need any context (e.g., repo, workflow, etc).
• Let’s you use OpenAI and other models, with a rate limit.

Save it into an environment variable, GITHUB_TOKEN

https://github.com/marketplace?type=models

Python Chatlas

import chatlas as clt
chat = clt.ChatGithub()
chat.chat("what is the capital of the moon?")

R Ellmer

library(ellmer)
chat <-  chat_github()
chat$chat("what is the capital of the moon?")