This is a Python script to clean and wrangle a messy dataset. The dataset contains 101 rows and 12 columns with encoding and delimiter errors, missing values, inconsistent and invalid data types, duplicated rows, invalid characters, and misspelt words.
This SQL file contains solutions to various SQL challenges from HackerRank. It includes solutions to 25 questions in the SQL (Basic)/(Intermediate) Select Skill in the Easy and Medium Difficulty category.
This Python script focuses on conducting exploratory data analysis (EDA) on a Prosper loan dataset. Despite potential questions about the data, my sole interest is in exploring the loan characteristics and factors that impact loan performance, which will be useful in gaining insights into those factors.
This SQL query creates six tables: Customers, LoanOfficers, Loans, LoanApplications, LoanPayments, and LoanRepayments. These tables are related to each other using foreign keys, ensuring referential integrity between them.
The Python script aims to predict whether a borrower's loan will be rolled based on their previous 6 months' financial records. The dataset had 5783 records and 11 fields, with null and duplicate values removed. Three classification models (Logistic Regression, Random Forest, and XGBoost) were tested, and XGBoost was recommended, providing an accuracy and precision score of 100% with false positives and false negatives of 0, after deriving debt-to-income ratio and interest rate to improve predictions.
The dataset was extracted from .csv files in the hotel system and included bookings, food orders, menu, requests, and rooms. SQL was used to conduct exploratory data analysis (EDA), and Power BI was used for visualization. The EDA process is contained in this SQL file.
Stakeholders are primarily interested in gaining insights, which are better understood through visual representations rather than raw numbers. To that end, I have invested significant time in studying the most suitable chart types for different audiences and specific insights that need to be conveyed. The links below provides access to my collection of dashboards featuring diverse data categories.