This SQL data cleaning project involved cleaning and transforming a messy dataset from an e-commerce company. The dataset contained missing values, duplicates, inconsistent data types, and other errors that needed to be fixed. The cleaning process involved various SQL queries and functions to handle the issues and transform the data into a more usable format. The resulting cleaned dataset was ready for analysis and provided valuable insights for the company.
The SQL code queries a COVID-19 database containing information about cases, deaths, vaccinations, population, and continent. The first set of queries retrieve data related to cases, deaths, and population percentages in Ghana and other countries. Another query looks at countries with the highest infection and death rates. There are also queries that group data by continent and look at the global data. Finally, the code joins the CovidDeaths and CovidVaccinations tables to analyze the correlation between population, vaccinations, and location
Through the use of Tableau, I have transformed data into powerful visualizations that provide insights and drive informed decision-making. My Tableau projects showcase my ability to explore, analyze, and present data in a meaningful way, allowing stakeholders to easily understand complex information and make data-driven decisions. I have used Tableau to build interactive dashboards and reports that help stakeholders understand trends, patterns, and insights hidden within data. This turns data into meaningful insights that drive decision-making.
This is a Python project that uses web scraping to monitor the price of a product on Amazon and sends email alerts when the price drops below a target price. It requires Python 3.x, BeautifulSoup, and Requests libraries to be installed. Users must set variables for the product URL, target price, email addresses, and password. The script can be run by navigating to the directory containing the script and running "python amazon_price_tracker.py". This project demonstrates the usefulness of Python for web scraping and automation.
This is a Python project that scrapes data from the CoinMarketCap API for the 15 best-performing cryptocurrencies and analyzes their price changes over time. The resulting data frame is then concatenated with the previously collected data frames (if any) and saved to a CSV file. Next, a loop is run 333 times with a 60-second sleep time in between. After the loop completes, the data is analyzed using pandas and seaborn to generate line plots for price changes of 15 selected cryptocurrencies over time.
This project is an Excel-based bike sales dashboard that analyzes average income per purchase based on gender, customer age brackets, and customer commute and purchase behavior. The dashboard includes dynamic interactive features such as filters and slicers, allowing users to explore the data in different ways and gain insights based on their own criteria. There are filters for marital status, region, education, house ownership, and car count. The analysis is performed using pivot tables, allowing for easy visualization and exploration of the data. The project demonstrates how Excel can be used as a powerful tool for data visualization and analysis, even for large datasets with complex relationships and multiple dimensions.
This project aims to predict the cost of SpaceX launches using machine learning algorithms and publicly available data collected from SpaceX's website and other sources. The data was cleaned, wrangled, and transformed into a suitable format for analysis, and exploratory data analysis was performed to identify patterns and correlations between different variables. The results of the project are presented in an interactive Plotly Dash dashboard that can be used to explore and compare the cost of SpaceX launches with other launch providers. The machine learning model, built using the scikit-learn library in Python, uses a decision tree algorithm to make predictions based on the data.
This project involved data cleaning and exploratory data analysis (EDA) of a California house pricing dataset, followed by building a house price prediction model using linear regression and decision tree regression algorithms. The models were evaluated using MSE, MAE, and R2 metrics. The decision tree regression model was fine-tuned using GridSearchCV to improve its performance. The performance of the tuned model was evaluated using the same metrics and compared to the linear regression model. This project highlights the importance of data cleaning, EDA, and hyperparameter tuning for accurate regression modeling. It is a valuable addition to a portfolio website for someone interested in data science or machine learning.
This project focuses on predicting customer churn using the telco-Customer-Churn dataset with 7043 rows and 20 columns. The target variable is "Churn," indicating whether the customer left within the last month. The dataset includes demographic, account, and service information. After cleaning, EDA is performed to visualize the data. Categorical values are mapped to numerical values using one-hot encoding. The data is standardized and split into training and testing sets. Three machine learning models, logistic regression, decision tree, and random forest, are trained and evaluated. The random forest model performed the best with an accuracy score of 78.1%.