How to Use AI Effectively

library(tidyverse)
library(elmer)


# {elmer} -----------------------------------------------------------------

chat <- chat_openai()

chat$chat("What is the easiest way to make a histogram in R?")

# System prompts with {elmer} ---------------------------------------------

chat <- chat_openai(
  system_prompt = "You are an expert R programmer who prefers the tidyverse."
)

chat$chat("What is the easiest way to make a histogram?")

# {pal} -------------------------------------------------------------------

library(pal)
library(ggpal2)

options(
  .pal_fn = "chat_openai"
)

ggplot(data, aes(x = variable)) +
geom_histogram(binwidth = width, fill = "lightblue", color = "black") +
  labs(title = "Histogram Title", 
       x = "X-axis Label", 
       y = "Frequency") +
  scale_fill_viridis_c()

tibble(data = c(1, 2, 2, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 5)) |>
  ggplot(aes(x = data)) +
  geom_histogram(binwidth = 1, 
                 fill = "lightblue", 
                 color = "black") +
  labs(title = "Histogram of Sample Data",
       x = "Values",
       y = "Frequency")
# This code is a small R script that creates a histogram using the `ggplot2` package. Let's go through it step by step:
# 
# 1. **Creating a tibble:**
#    - `tibble(data = c(1, 2, 2, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 5))`: This line creates a `tibble`, which is a modern version of a data frame provided by the `tibble` package. It is essentially used for storing data in a tabular format.
#    - The `tibble` here contains a single column named `data` with the specified sequence of numbers.
# 
# 2. **Pipe operator (|>):**
#    - The pipe operator `|>` (introduced in R 4.1.0) passes the `tibble` created in the first step as input to the following function. This helps streamline the operations and makes the code more readable.
# 
# 3. **Creating a ggplot2 plot:**
#    - `ggplot(aes(x = data))`: This initializes a `ggplot` object and specifies that the `x`-axis aesthetic will use the `data` column from the `tibble`.
# 
# 4. **Adding a histogram layer:**
#    - `geom_histogram(...)`: This adds a histogram to the plot with the following specifications:
#      - `binwidth = 1`: This sets the width of each bin (interval) to 1. It will group all the values within each unit into a single bin.
#      - `fill = "lightblue"`: This specifies that the bars of the histogram should be filled with a light blue color.
#      - `color = "black"`: This outlines the bars with a black color.
# 
# 5. **Adding labels:**
#    - `labs(...)`: This adds labels to the plot to enhance readability.
#      - `title = "Histogram of Sample Data"`: Sets the main title of the plot.
#      - `x = "Values"`: Labels the x-axis with the name "Values".
#      - `y = "Frequency"`: Labels the y-axis with the name "Frequency", indicating that the y-axis represents the frequency of each value.
# 
# In summary, the code creates a histogram that visualizes the frequency of each number in the given `data` vector. The plot will have bars colored light blue


# Make my own pal ---------------------------------------------------------

prompt_new(
  "explaincode",
  interface = "suffix"
)

directory_load()

prompt_new(
  "debugcode",
  interface = "suffix"
)

tibble(data = c(1, 2, 2, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 5)) |>
  ggplot(aes(x = data)) +
  geom_histogam(binwidth = 1, 
                 fill = "lightblue", 
                 color = "black") +
  labs(title = "Histogram of Sample Data",
       x = "Values",
       y = "Frequency")



# Caveat ------------------------------------------------------------------

# Write code in tidycensus to bring in the percentage of households below 200% poverty level in each census tract in Michigan

library(tidycensus)

mi_poverty <-
  get_acs(
    geography = "tract",
    table = c(
      below_200 = "B17002_002", # Count below 200% poverty
      total = "B17002_001" # Total population for poverty status
    ),
    state = "MI",
    year = 2022,
    geometry = FALSE
  ) |>
  # Pivot wider to get columns side by side
  pivot_wider(
    names_from = "variable",
    values_from = "estimate"
  ) |>
  # Calculate percentage
  mutate(
    pct_below_200 = (below_200 / total) * 100
  ) |>
  # Clean up the data
  select(GEOID, NAME, pct_below_200)