Digital Shelf Design and Subsequent Consumer Response in E-Commerce: A Multi-Objective Evolutionary Product Assortment Analysis Using Amazon Books Data.

Repository Overview

This repository contains the R code used in the study titled “Digital Shelf Design and Subsequent Consumer Response in E-Commerce: A Multi-Objective Evolutionary Product Assortment Analysis Using Amazon Books Data.”

The repository includes 15 files in total: 14 R scripts and 1 Markdown file. Together, these files cover the full analytical workflow, from raw data processing to model comparison, backtesting, and figure generation.

The code was developed and run in R version 4.5.2.

All figures reported in the manuscript were produced in R.

File structure and functions

1. 01_books_review_chunk_processing.R  
   Processes the raw Books review data in chunks and prepares the initial review-level analytical structure.

2. 02_meta_books_cleaning_and_categor...  
   Cleans the meta_Books file and standardizes category-related metadata fields.

3. 03_review_metadata_merge_and_tem...  
   Merges review data with metadata and prepares the temporal analytical structure used in the study.

4. 04_sample_screening_and_analysis_da...  
   Screens the sample and constructs the main analysis dataset after filtering and eligibility checks.

5. 05_score_component_construction.R  
   Constructs the main score components used in the study, including quality signals, social proof, diversity, and discovery.

6. 06_candidate_pool_screening.R  
   Screens candidate products for the exact-K assortment procedure and prepares the category-level candidate pools.

7. 07_baseline_assortment_models.R  
   Runs the baseline assortment models used for comparison, including rating-only, popularity-only, and hybrid specifications.

8. 08_exact_k_genetic_assortment_optim...  
   Implements the genetic algorithm used for exact-K assortment optimization.

9. 09_exact_k_solution_repair.R  
   Applies the deterministic repair logic to ensure that each final solution satisfies the exact-K constraint.

10. 10_random_baseline_replications.R  
    Produces repeated random-baseline runs used as a comparison benchmark.

11. 11_market_response_backtesting.R  
    Evaluates subsequent-period market response and performs the backtesting stage.

12. 12_model_comparison_and_rank_gap...  
    Compares model outputs and computes rank-gap patterns across evaluation criteria.

13. 13_descriptive_statistics_and_summar...  
    Produces descriptive statistics and summary outputs for the final analytical sample.

14. 14_figure_generation_and_export.R  
    Generates and exports the figures reported in the manuscript.

15. RFADMF KOD HARITASI.md  
    Provides a roadmap of the coding workflow and explains how the scripts relate to one another.

Important note

The scripts are shared for transparency and reproducibility. Users may need to adapt file paths to their local environment before running the code. The repository assumes access to the underlying dataset(s) used in the study.