zomato_data_analysis

The project focuses on data cleaning, transformation, and exploratory data analysis (EDA) using Python. this is purely data-focused, aimed at uncovering patterns in ratings, cost, and services like online ordering and table booking.

---

📊 Project Objectives

• Clean and preprocess the raw Zomato dataset.
• Explore the data visually to find patterns and trends.
• Understand how different features like online_order, book_table, rest_type, and city affect ratings and popularity.

---

🔧 Data Cleaning Steps Performed

1. Removed unnecessary columns like url, phone, address, etc.
2. Replaced missing and inconsistent values in:
   • rate (e.g., "NEW", "-") replaced with numeric mode.
   • approx_cost(for two people) cleaned and converted to integers.
   • rest_type and cuisines filled with mode values.
3. Removed duplicate entries.
4. Cleaned text-based noise in restaurant names.
5. Converted categorical columns like online_order and book_table to binary (Yes/No → 1/0).
6. Extracted and analyzed frequent dishes mentioned in dish_liked.

---

📈 EDA Highlights

Some questions we explored through visualizations:

• How does online ordering affect restaurant ratings?
• Do restaurants that allow table booking have higher ratings?
• Which areas (city) receive the most customer votes?
• What are the most common restaurant types?
• How does average cost relate to ratings?

All charts are stored in the /charts folder and generated via Seaborn and Matplotlib.

---

📁 Dataset Info

The original dataset was sourced from [Zomato's public listing in Bangalore] and contains ~43,000 rows with the following attributes:

• Restaurant name
• Location
• Online ordering (Yes/No)
• Table booking availability
• Average rating
• Cuisines served
• Cost for two
• Votes received
• And more...

---

🛠️ Tools & Libraries Used

• Python 3.x
• Pandas for data manipulation
• NumPy for numerical handling
• Matplotlib & Seaborn for visualization
• Jupyter Notebook for interactive development

---