website·Indexed 30 days ago
https://laresbernardo.github.io/lares/index.htmlAn R package providing a diverse set of utilities for data science, including H2O AutoML integration, text mining (RAKE, sentiment breakdown, wordclouds), financial portfolio reporting, AES-256-CBC file encryption, and specialized tools for solving Sudoku and mazes.
lares library is an R package developed to provide tools for better and faster analytics, visualization, data mining, and machine learning tasks. It serves as a sidekick for Lean Analytics and Robust Exploration.lares is an R package designed to automate and accelerate analytics and machine learning workflows. Its core capabilities include:
ohse()).freqs(), distr(), corr_var(), corr_cross()).fb_*), and OpenAI API (gpt_*).read_encrypted(), write_encrypted()).lares is an R package designed to automate and accelerate analytics and machine learning workflows. It provides specialized function families for:
clusterKmeans function automates the process of grouping data into distinct non-overlapping subgroups. It handles preprocessing steps such as one-hot encoding for categorical variables and data normalization. Note that K-Means assumes spherical cluster shapes and may not perform well with elliptical clusters. Scaling/standardizing data is recommended.# Example: Run K-Means with a specific number of clusters (k=3)
clusters <- clusterKmeans(df, k = 3)
# Access the resulting dataframe with the 'cluster' column
clustered_df <- clusters$df
# Plot cross-correlations for each cluster
plot(clusters$correlations)write_encrypted() to write a list, vector, or any JSON-able object directly into an encrypted file. If no key is provided, the function generates one and returns it.my_secrets <- list(
api_key = "some_secret_api_key_123",
username = "data_analyst",
server = "production.server.com"
)
# Write to file and capture the generated key
raw_key <- write_encrypted(data = my_secrets, output_file = "secrets.enc", quiet = FALSE)mplot_gain function creates a cumulative gains chart (or 'gains plot'). This visualization helps determine the expected target capture rate when selecting the best X deciles of a population based on model scores. It shows the percentage of total positive cases 'gained' relative to the percentage of the total population targeted.# Example for a Binomial Model
mplot_gain(
tag = dfr$class2$tag,
score = dfr$class2$scores,
caption = "Titanic Survived Model",
target = "TRUE"
)balance_data to achieve a 1:1 ratio or a specific custom ratio using the Titanic dataset (dft).# Load Titanic dataset
data(dft)
# Balance to a 1:1 ratio (rate = 1)
# This will automatically reduce the most frequent label
df_balanced <- balance_data(dft, Survived, rate = 1)
# Balance to a 0.5 ratio, specifically reducing the "TRUE" label
df_custom <- balance_data(dft, .data$Survived, rate = 0.5, target = "TRUE")ohse function automatically transforms a dataframe by converting categorical columns into numerical ones using one-hot encoding. It can also generate additional features from date/time columns and include holiday data.ohse(
df,
redundant = FALSE,
drop = TRUE,
ignore = NULL,
dates = FALSE,
holidays = FALSE,
country = "Venezuela",
currency_pair = NA,
trim = 0,
limit = 10,
variance = 0.9,
other_label = "OTHER",
sep = "_",
quiet = FALSE,
...
)try_require to manage library dependencies in R scripts. Set stop = TRUE for mandatory dependencies that should halt the script, and stop = FALSE for optional packages where a warning is sufficient to notify the user.# Mandatory dependency
try_require("ggplot2", stop = TRUE)
# Optional dependency
try_require("shiny", stop = FALSE)token <- "YOURTOKEN"
which <- "act_20846447"
insights_adset <- fb_insights(
token, which,
time_increment = "1",
report_level = "campaign",
fields = c(
"adset_id",
"reach",
"frequency",
"spend",
"cpm",
"objective",
"optimization_goal"
)
)The lares package provides shorthand functions to quickly format axis text in ggplot2. These functions act as wrappers around continuous scales to apply specific formatting styles:
scale_x_comma() / scale_y_comma(): Formats numbers with commas.scale_x_percent() / scale_y_percent(): Formats numbers as percentages.scale_x_dollar() / scale_y_dollar(): Formats numbers as currency (dollars).scale_x_abbr() / scale_y_abbr(): Formats numbers using abbreviations.library(ggplot2)
# Example using dollar and abbreviated scales
ggplot(df, aes(x = sales, y = volume)) +
geom_point() +
scale_x_dollar() +
scale_y_abbr()
# Example using comma and percent scales with standard ggplot2 scale arguments (e.g., limits)
ggplot(df, aes(x = listings, y = log(inventory))) +
geom_point() +
scale_x_comma() +
scale_y_percent(limits = c(0, 3))encrypt_file() to encrypt a file. The process pads the data, generates a random Initialization Vector (IV), and prepends the IV to the encrypted data in the output file. The encryption key must be a raw vector of 32 bytes (256 bits).# Generate a random 32-byte key
raw_key <- openssl::rand_bytes(32)
# Encrypt the file
encrypt_file(input_file = "secrets.txt", output_file = "secrets.enc", key = raw_key)