lares

Lean Analytics and Robust Exploration Sidekick

lares is an R package designed to automate, improve, and accelerate everyday analytics and machine learning tasks. It offers a wide variety of functions grouped in families for:

Machine Learning: Streamlined model training and evaluation, including friendly AutoML pipelines.
Data Cleaning & Processing: Functions to quickly prepare your data for modeling or analyzes.
Exploratory Data Analysis (EDA): Instantly visualize and summarize your data.
Reporting: Easily generate comprehensive reports to share insights for MMM and ML models.
Visualization: Out-of-the-box plotting for classification and regression models, timelines, and more.
API Integrations & Scrapers: Simplify data collection from various sources.
Time Series & Portfolio Analysis: Specialized utilities for financial and temporal data.
Credentials & Secrets Management: Securely handle sensitive information in your analytics pipelines.
NLP & Text Analytics: Tools to analyze and process text data.

Tip: See all available functions and documentation here or type ?lares:: in RStudio to explore interactively.

Installation

# CRAN VERSION
install.packages("lares")

# DEV VERSION (latest updates)
# If you don't have remotes installed yet, run: install.packages('remotes')
remotes::install_github("laresbernardo/lares")
# For a full installation with recommended dependencies:
remotes::install_github("laresbernardo/lares", dependencies = TRUE)

Windows users: You may need to install RTools to build the dev version.

Read about `lares` in action!

AutoML Quickstart: Introduction to AutoML using lares
Model Results Visualization: Classification | Regression
Marketing Mix Model Selection: Select the right MMM candidate
Cross-Correlations: Find Insights with Ranked Cross-Correlations
Secure Credentials: Manage Secrets in R
Portfolio Analysis: Performance & Reporting
Fun games: Wordle, Scrabble, Sudoku, Mazes, etc.
More Examples: Read other posts

Popular Functions

h2o_automl(), plot_model_results() – Automated machine learning pipeline with optimal model selection and visualizations.
freqs(), distr(), corr_var(), corr_cross() – Instantly summarize, visualize, and uncover relationships in your data.
ohse() – Efficient and smart one-hot encoding for categorical variables.
cache_* – Speed up workflows by caching expensive computations.
robyn_* – Additional functions to support Robyn inputs and outputs.
fb_* – Interact with Meta’s Marketing API
gpt_* – Structured prompts builder and interact with OpenAI’s API
read_encrypted(), write_encrypted() – Interact with encrypted files to keep secrets safe
…and many more!

Getting Started & Help