title	AI dan AR dalam Gamifikasi SIAR Halal apps: Katalisasi Pertumbuhan Sosial Ekonomi dalam Kerangka Digital Halal
author	irhafidz, nurainir, hadziq, Siska Arifiani
date	13/03/2025
output	raw_data (moneyspent_cleaned.csv) cleaned_data (out.csv), sentiment_siarhalal.ipynb

Repository Structure

📂 Sentiment-Analysis-SIARHalal-MSMEs
 ├── data/                # Raw and processed datasets
 │   ├── moneyspent_cleaned.csv      # Original dataset
 │   ├── out.csv    # Processed dataset with sentiment scores
 │
 ├── notebooks/           # Jupyter Notebooks for analysis
 │   ├── sentimentsiarhalal.ipynb  # Main notebook
 │
 ├── README.md            # Project introduction (this file)
 ├── LICENSE              # License information

Introduction

This repository is part of the research project "AI dan AR dalam Gamifikasi SIAR Halal apps: Katalisasi Pertumbuhan Sosial Ekonomi dalam Kerangka Digital Halal", which is a recipient of RGBI 2024 (Research Grant Bank Indonesia) under TIM ID 2516 – TOPIK 1: Pemanfaatan AI untuk meningkatkan Inklusi Ekonomi & Keuangan Konvensional dan Syariah.

Research Theme:

• Sub-tema 3: Pemanfaatan AI untuk Mendorong Inklusi Ekonomi khususnya kepada pelaku ekonomi kecil dan menengah

This project focuses on leveraging AI and AR technologies to enhance the gamification experience in SIAR Halal apps, aiming to boost digital halal economy growth and improve economic inclusion, particularly for small and medium enterprises (MSMEs).

Key features of this project include:

• Preprocessing user reviews (cleaning, stopword removal, tokenization)
• Sentiment analysis using TextBlob (polarity and subjectivity scoring)
• Data aggregation and visualization (analyzing sentiment trends per food stall)

The primary goal is to identify trends in customer feedback, highlight areas for business improvement, and understand how sentiment correlates with ratings and customer visits.

Dataset

The dataset consists of user-generated reviews for multiple food stalls, including:

• msme: Name of the food stall
• Number_of_Visits: Total visits recorded
• Number_of_Feedbacks: Number of customer reviews
• Average_Rating: Average rating score
• Average_Spend: Average spending per customer
• Sentiment: Sentiment polarity score (-1 = Negative, 0 = Neutral, 1 = Positive)

Methodology

The sentiment analysis pipeline follows these steps:

1. Data Preprocessing:
   • Tokenization
   • Stopword removal using Sastrawi
   • Normalization (removing special characters, lowercasing, stemming, lemmatization)
2. Sentiment Analysis:
   • TextBlob to compute polarity and subjectivity scores
   • Aggregation of sentiment scores per food stall
3. Exploratory Data Analysis (EDA):
   • Word cloud visualization
   • Sentiment distribution charts
   • Correlation analysis between sentiment, ratings, and visits