Using gander and ellmer – AI-Usage-in-R

Introduction

I am going to use gander to create a simple example of how to use it. It seems gander can allow the user to see the data they are working with and is able to generate accurate R code for graphs.

Setting Up R Packages

library(tidyverse)
library(ggformula)
library(gander)
library(ellmer)

Setting up `gander` and `ellmer`

There are a few steps involved in setting up gander and ellmer to work with your preferred LLM (Large Language Model) service. The steps are as follows:

Install the gander and ellmer packages.
Get your API Key, from say Anthropic, ChatGPT, or Ollama.

Anthropic: visit https://console.anthropic.com/settings/keys
ChatGPT: visit https://platform.openai.com/api-keys
Ollama: visit https://ollama.com/ # runs LLMs LOCALLY! NO API Key needed.

See “Choosing a model” in vignette(“gander”, package = “gander”) to learn more.

Save the API_KEY of you preferred LLM service in your .Rprofile file. For example, if you are using Ollama, you would add the following line to your .Rprofile, as follows:

## RUN THESE IN YOUR CONSOLE
install.packages("usethis")
library(usethis)
usethis::edit_r_profile()

This will open the .Rprofile file in your RStudio editor. You will need to add (one of ) the following lines to the file:

OPENAI_API_KEY = “YOUR_API_KEY_HERE”

ANTHROPIC_API_KEY = “YOUR_API_KEY_HERE”

etc.

Also add one of these line to the .Rprofile file:

options(.gander_chat = ellmer::chat_anthropic())

If you wish to run models locally for free, you can use Ollama. In that case, you will need to:

install the Ollama CLI https://ollama.com/download
download a model from https://ollama.com/models, say “llama3.1”
then add the following line to your .Rprofile file: options(.gander_chat = ellmer::chat_ollama(model = "llama3.1"))
MacOS: Ensure you start ollama server by running ollama serve in your terminal.
Windows: Add ollama.exe to your startup programs

Setup the gander assistant: The gander assistant is interfaced with the via the gander addin. For easiest access, we recommend registering the gander addin to a keyboard shortcut.

In RStudio, navigate to Tools > Modify Keyboard Shortcuts > Search "gander". The package author suggests Ctrl+Alt+G (or Ctrl+Cmd+G on macOS).

Restart RStudio to ensure the changes take effect.

Using `gander`

The gander package speaks with ellmer to “speak with an LLM”. Hence there is a small amount of setup required for gander to deliver its goods.

As per @ref1, let us use the stackoverflow data from the modeldata package to create our graphs.

data("stackoverflow", package = "modeldata")
stackoverflow

# A tibble: 5,594 × 21
   Country        Salary YearsCodedJob OpenSource Hobby CompanySizeNumber Remote
   <fct>           <dbl>         <int>      <dbl> <dbl>             <dbl> <fct> 
 1 United Kingdom 1   e5            20          0     1              5000 Remote
 2 United States  1.3 e5            20          1     1              1000 Remote
 3 United States  1.75e5            16          0     1             10000 Not r…
 4 Germany        6.45e4             4          0     0              1000 Not r…
 5 India          6.64e3             1          0     1              5000 Not r…
 6 United States  6.5 e4             1          0     1                20 Not r…
 7 United States  1.10e5            13          0     1                20 Not r…
 8 Germany        5.38e4             4          1     0              5000 Not r…
 9 United States  1.09e5             7          1     1                20 Not r…
10 Germany        8.39e4            17          1     1                20 Not r…
# ℹ 5,584 more rows
# ℹ 14 more variables: CareerSatisfaction <int>, Data_scientist <dbl>,
#   Database_administrator <dbl>, Desktop_applications_developer <dbl>,
#   Developer_with_stats_math_background <dbl>, DevOps <dbl>,
#   Embedded_developer <dbl>, Graphic_designer <dbl>,
#   Graphics_programming <dbl>, Machine_learning_specialist <dbl>,
#   Mobile_developer <dbl>, Quality_assurance_engineer <dbl>, …

names(stackoverflow)

 [1] "Country"                             
 [2] "Salary"                              
 [3] "YearsCodedJob"                       
 [4] "OpenSource"                          
 [5] "Hobby"                               
 [6] "CompanySizeNumber"                   
 [7] "Remote"                              
 [8] "CareerSatisfaction"                  
 [9] "Data_scientist"                      
[10] "Database_administrator"              
[11] "Desktop_applications_developer"      
[12] "Developer_with_stats_math_background"
[13] "DevOps"                              
[14] "Embedded_developer"                  
[15] "Graphic_designer"                    
[16] "Graphics_programming"                
[17] "Machine_learning_specialist"         
[18] "Mobile_developer"                    
[19] "Quality_assurance_engineer"          
[20] "Systems_administrator"               
[21] "Web_developer"

Let us now create a simple bar plot based on the variable “Country”.

Scatter plot for Salary vs Years coded

gf_point(data = stackoverflow, Salary ~ YearsCodedJob) %>%
  gf_lm() %>% 
  gf_theme(theme_bw())

Warning: Using the `size` aesthetic with geom_line was deprecated in ggplot2 3.4.0.
ℹ Please use the `linewidth` aesthetic instead.

gf_boxplot(Salary ~ YearsCodedJob, group = ~ YearsCodedJob,data = stackoverflow)

Introduction

Setting Up R Packages

Setting up gander and ellmer

Using gander

Scatter plot for Salary vs Years coded

References

Setting up `gander` and `ellmer`

Using `gander`