Data-Analysis-Project-MySQL-Python-PowerBi

A comprehensive end-to-end data analysis project that demonstrates the complete data pipeline: from database creation and data cleaning to machine learning and visualization. This project analyzes customer transaction data using MySQL, Python, and Power BI.

🎯 Objectives

1. Create and populate a MySQL database with transaction data
2. Clean and preprocess messy data using Python and Pandas
3. Perform exploratory data analysis
4. Build predictive models with scikit-learn
5. Visualize insights using Power BI dashboards

1) Create Database (MySQL)

Data Quality Issues (Intentional for Cleaning Practice)

1. Duplicate records
2. Special characters in names (?, ., /, _, 11)
3. Inconsistent capitalization
4. Missing values in critical fields
5. Typos (e.g., "expence" instead of "expense")

2) Data Cleaning and Regression (Python)

Step 1: Import the Database into Python

Step 2: Data cleaning

✅ Remove Duplicates

✅ Fix our columns (first and last names) using strip

✅ Fix the transaction_type and category columns using "replace"

✅ Fill in the blancks/Null values

✅ Drop down the rows without any currency

✅ Spelling correction

🧾 Raw Dataset (Before Cleaning)

📋 Dataset after Cleaning

Step 3: Random Forest (Machine Learning)

Classification Model

1. Algorithm: Random Forest Classifier
2. Target Variable: Category (bonus, rent, utilities)
3. Features: Amount, transaction_type, currency
4. Train/Test Split: 80/20

Predictions

1) Next 2 categories - utilities

2) Probability of this prediction - 78% (utilities)

3) Probability of another prediction:

bonus - 3%

rent - 14%

4) Accuracy: 0.5

3) 📈 Dashboard (Power Bi)

Key Metrics

1. Total Amount: €5,020
2. Year: 2024
3. Standard Deviation: 397
4. Total Customers: 9

Visualizations

Amount by Quarter - Bar chart showing transaction trends

Category Distribution - Pie chart breakdown

Customer Details - Interactive table with filters

Transaction Type Filter - Income vs Expense analysis

📊 Key Insights

Transaction Distribution

Bonuses account for nearly half of all transactions (44.58%)

Utilities represent a third of transactions (33.33%)

Rent payments are the smallest category (22.08%)

Quarterly Trends

Q1 and Q2 show highest transaction volumes

Q3 has significantly lower activity

Q4 shows moderate recovery

Currency Usage

EUR is the dominant currency (7 out of 9 customers)

USD is used by 2 customers

🛠️ Technologies Used

Database

MySQL 8.0 - Relational database management

Data Processing & Analysis

Python 3.11

pandas - Data manipulation and cleaning

numpy - Numerical computations

scikit-learn - Machine learning models

Business Intelligence

Power BI Desktop - Interactive dashboards