🚗 Baly User Experience Analysis

A data-driven sentiment analysis of Baly — Iraq's leading ride-sharing and food delivery platform — examining the gap between user expectations and driver realities.

---

📊 Interactive Dashboard

🔍 What We Did

1. Manual Data Collection

• 280 real user reviews from Google Play (Baly User App)
• 100 real driver reviews from Google Play (Baly Captain App)
• Collected over one week by reading and categorizing each review individually
• No datasets, no shortcuts — real data from a real app

2. Manual Categorization & Labeling

Each review was tagged with:

• Category — What is the complaint about?
• Rating — Star rating (1-5)
• Helpful_Count — How many users found it helpful
• Date — When was it posted
• Gender — User/Driver gender (M/F)

3. Deep Analysis & Insights

Discovered hidden patterns by comparing the two datasets side-by-side.

---

💡 Key Findings

Finding #1: The Wage Crisis

Driver Reviews Analysis:
99% of complaints → "Wages are too low"
Only 1% → App bugs, support issues

Implication:
The platform is technically sound for drivers,
but economically unsustainable.

Finding #2: The Rating Contradiction

Same app, different ratings:

Mobile (Google Play - Iraq only)  → 4.2 ⭐
Web (Google Play - Global)        → 3.6 ⭐

Why? Geographic filtering + limited reviews inflate mobile score.

Finding #3: The Two-Sided Gap

USERS complain about:
├─ Price (25%)
├─ App bugs (22%)
├─ Driver behavior (14%)
└─ Other (39%)

DRIVERS complain about:
├─ Low wages (99%)
└─ Other (1%)

Same platform = completely different pain points

Finding #4: The Silent Majority

Users who write reviews are angry.
Users who stay silent are neutral or happy.

The 4.2 rating on mobile hides a silent approval gap.
Real sentiment may be lower than displayed.

---

📁 Dataset

Users Dataset (baly_users.xlsx)

• 280 reviews
• Categories: Price_Issue, App_Technical, Support_Service, Positive
• Date Range: 2022–2026
• Manual verification: 100%

Drivers Dataset (captain_baly.xlsx)

• 100 reviews
• Categories: Payment_Issue, App_Technical, Trip_Cancellation, Support_Service, Positive
• Date Range: 2022–2026
• Manual verification: 100%

---

📊 Dashboard Overview

The interactive dashboard reveals:

• Timeline visualization of reviews over time
• Rating distribution comparison (Users vs Drivers)
• Category breakdown with donut charts
• Trend analysis showing if app improved over time
• Key statistics and insights cards

Open the dashboard above to explore the data!

---

🎯 Why This Matters

For Baly (Internal)

• Users expect service quality; drivers expect fair pay
• The app isn't broken for drivers — the business model is
• Low ratings don't reflect app quality; they reflect wage concerns

For Data Science Community

• Rare example of multi-stakeholder sentiment analysis
• Manual data collection from live platform (not Kaggle)
• Local dataset — addresses gap in Middle Eastern app analysis

For Job Seekers / Portfolio

• Demonstrates ability to collect real data manually
• Shows comparative analysis skills
• Reveals business insights from raw reviews
• Proves attention to data quality and context

---

🚀 How to Use

1. View the Dashboard

Click the button above — no installation needed.

2. Analyze the Data

import pandas as pd

users = pd.read_excel("baly_users.xlsx")
drivers = pd.read_excel("captain_baly.xlsx")

# Compare wage complaints between drivers
wage_complaints = drivers[drivers['Category'] == 'Payment_Issue']
print(f"Wage complaints: {len(wage_complaints) / len(drivers) * 100:.1f}%")

3. Extend the Analysis

Add your own visualizations, NLP sentiment analysis, or predictive models.

---

📊 Key Statistics

Total Reviews          : 380
Users                  : 280 reviews
Drivers                : 100 reviews
Coverage               : 4 years (2022–2026)
Data Quality           : 100% manual verification
Geographic Coverage    : Iraq (Baghdad-focused)

---

⚠️ Data Limitations

Strengths ✅

• Real data from active Google Play reviews
• Manual categorization ensures accuracy
• Complete dataset — no missing values
• Verified sources — all from official Baly apps

Limitations ⚠️

• Sample size for drivers — only 100 reviews (limited generalization)
• Geographic bias — mostly Baghdad-based users
• Recency bias — most reviews from 2025–2026
• Selection bias — only people motivated to review (usually angry or very happy)
• Language — Arabic reviews; translations may vary
• Platform limitation — doesn't capture in-app feedback or customer support tickets

---

👤 Author

Abbas Hussein
Data Analyst | Python | Business Intelligence
LinkedIn · GitHub

---

🔮 Future Work

Potential expansions:

• NLP Sentiment Analysis — Beyond manual categorization
• Topic Modeling — Discover hidden complaint patterns
• Time Series Forecasting — Will complaints improve?
• Predictive Models — What drives users to leave reviews?

---

Last Updated: June 2026
Real data. Real insights. No shortcuts.