lares

🥈Silver

Lares is an R-based AI skill for marketing analytics and machine learning. It automates data exploration, predictive modeling, and marketing mix modeling. Connects to H2O and Robyn for scalable machine learning. Ideal for marketing teams needing automated insights and predictive analytics.

239500Updated 3d ago

Intermediate30min to implementmarketing

Saves ~240 min per use

Quick InstallView Source

git clone https://github.com/laresbernardo/lares.git

Works with:

Claude

Overview

About This Skill

How to Use

1. **Prepare Your Data:** Export your marketing dataset (spend, conversions, customer attributes) in CSV/Excel format. Ensure columns are labeled clearly (e.g., 'channel', 'spend', 'revenue', 'date'). Clean missing values and outliers using tools like Excel or Python (Pandas). 2. **Set Up Lares:** Install the Lares package in R (`install.packages('lares')`) and load it (`library(lares)`). Connect your dataset to H2O for scalable modeling (`h2o.init()`). If using Robyn for MMM, ensure it’s installed (`devtools::install_github('facebookexperimental/Robyn')`). 3. **Run the Analysis:** Use the prompt template to generate the analysis. Replace [MARKETING_DATASET] with your data file path, [MARKETING_METRIC] with your target (e.g., 'revenue'), and [MARKETING_CHANNELS] with your channels (e.g., 'paid_search, email, social'). 4. **Review Outputs:** Examine the EDA insights, model performance metrics (R², MAE), and budget recommendations. Focus on channels with high feature importance and ROI gaps. 5. **Implement Changes:** Apply the recommended budget shifts in your marketing platforms (e.g., Google Ads, Meta Ads Manager). Set up a dashboard (e.g., Tableau, Power BI) to track KPIs weekly and retrain the model monthly.

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/laresbernardo/lares

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Use Lares to analyze [MARKETING_DATASET] and generate insights. Perform exploratory data analysis to identify key trends, then build a predictive model to forecast [MARKETING_METRIC] (e.g., sales, conversion rates). Compare the model's performance against a baseline and suggest optimal budget allocation across [MARKETING_CHANNELS] based on the results. Include visualizations of feature importance and model predictions.

Example Output

### Lares Marketing Analytics Report

**Dataset Overview:** The analysis used a 12-month dataset from a mid-sized e-commerce retailer, including 500K+ transactions, 15 marketing channels (paid search, social, email, etc.), and 20+ customer attributes (demographics, past behavior). Key metrics tracked: daily spend, impressions, clicks, conversions, and revenue.

**Exploratory Data Analysis:**
The initial exploration revealed several critical insights:
- **Channel Performance:** Paid search (Google Ads) drove 42% of total conversions but only 35% of the budget, indicating high efficiency. Social media (Meta/Instagram) had the lowest ROI (ROAS of 1.8 vs. industry average of 2.5).
- **Seasonality:** Revenue spiked by 35% during Q4 holiday campaigns, with email marketing contributing 28% of total holiday sales.
- **Customer Segments:** High-value customers (LTV > $500) responded best to personalized email campaigns (CTR 8.2%) and retargeting ads (CTR 6.5%).

**Predictive Modeling:**
A Gradient Boosting Machine (XGBoost) model was trained to predict daily revenue based on marketing spend across channels. The model achieved an R² of 0.87 and a MAE of $1,200 (vs. baseline MAE of $3,800). Feature importance analysis showed:
- **Top Drivers:** Paid search spend (32%), email campaign frequency (22%), and retargeting ad impressions (18%).
- **Underperformers:** Display ads and influencer partnerships had negligible impact on revenue.

**Budget Optimization:**
Using Robyn (an MMM tool integrated with Lares), we simulated reallocating the $500K monthly budget based on the model’s predictions:
- **Recommended Allocation:**
  - Paid Search: +15% ($180K → $207K)
  - Email Marketing: +10% ($90K → $99K)
  - Social Media: -20% ($80K → $64K)
  - Display Ads: -50% ($50K → $25K)
- **Projected Impact:** Expected revenue increase of 18% ($1.2M → $1.42M/month) with a 12% reduction in CAC.

**Visualizations:**
1. **Channel ROI Heatmap:** Highlighted Paid Search and Email as top performers.
2. **Feature Importance Bar Chart:** Showed spend and frequency as dominant predictors.
3. **Budget Allocation Sankey Diagram:** Illustrated the shift from underperforming to high-impact channels.

**Next Steps:**
- Implement the recommended budget reallocation in Google Ads and the email platform.
- Monitor model performance weekly and retrain monthly to account for seasonality.
- Test a 10% increase in retargeting ad spend to validate its impact on high-LTV customers.