Dataframe Manipulation

Download or read book DATAFRAME MANIPULATION: THEORY AND APPLICATIONS WITH PYTHON AND TKINTER written by Vivian Siahaan and published by BALIGE PUBLISHING. This book was released on 2024-08-13 with total page 431 pages. Available in PDF, EPUB and Kindle. Book excerpt: A DataFrame is a fundamental data structure in pandas, a powerful Python library for data manipulation and analysis, designed to handle two-dimensional, labeled data akin to a spreadsheet or SQL table. It simplifies working with tabular data by supporting various operations like filtering, sorting, grouping, and aggregating. DataFrames are easily created from lists, dictionaries, or NumPy arrays and offer flexible data handling, including managing missing values and performing input/output operations with different file formats. Key features include hierarchical indexing for multi-level grouping, time series functionality, and integration with libraries such as NumPy and Matplotlib. DataFrame manipulation encompasses filtering, sorting, merging, grouping, pivoting, and reshaping data, while also allowing custom functions, handling missing data, and managing data types. Mastering these techniques is crucial for efficient data analysis, ensuring clean, transformed data ready for deeper insights and decision-making. In chapter 2, in the first project, we filter a DataFrame named employee_data, which includes columns like 'Name', 'Department', 'Age', 'Salary', and 'Years_Worked', to find employees in the 'Engineering' department with a salary exceeding $70,000. We create the DataFrame using sample data and apply boolean indexing to achieve this. The boolean masks employee_data['Department'] == 'Engineering' and employee_data['Salary'] > 70000 identify rows meeting each condition. Combining these masks with the & operator filters the DataFrame to include only those rows where both conditions are met, resulting in a subset of employees who fit the criteria. The final output displays this filtered DataFrame. In second project, we filter a DataFrame named sales_data, which includes columns such as 'Product', 'Category', 'Quantity Sold', 'Unit Price', and 'Total Revenue', to find products in the 'Electronics' category with quantities sold exceeding 100. We use boolean indexing to achieve this: sales_data['Category'] == 'Electronics' creates a mask for rows in the 'Electronics' category, while sales_data['Quantity_Sold'] > 100 identifies rows where quantities sold are above 100. By combining these masks with the & operator, we filter the DataFrame to include only rows meeting both conditions. The final output displays this filtered subset of products. In third project, we filter a DataFrame named movie_data, which includes columns such as 'Title', 'Genre', 'Release Year', 'Rating', and 'Box Office Earnings', to find movies released after 2010 with a rating above 8. We use boolean indexing where movie_data['Release_Year'] > 2010 creates a mask for movies released after 2010, and movie_data['Rating'] > 8 identifies movies with ratings higher than 8. By combining these masks with the & operator, we filter the DataFrame to include only the rows meeting both conditions. The final output displays the subset of movies that fit these criteria. The fourth project demonstrates a Tkinter-based GUI application for filtering a sales dataset using Python libraries Tkinter, Pandas, and PandasTable. The application allows users to interact with a table displaying sales data, applying filters based on product category and quantity sold. The filter_data() function updates the table to show only items from the selected category with quantities exceeding the specified value, while the refresh_data() function resets the table to display the original dataset. The GUI includes input fields for category selection and quantity entry, along with buttons for filtering and refreshing. The sales data is initially presented in a PandasTable with a toolbar and status bar. Users interact with the interface, which updates and displays filtered data or the full dataset as needed. The fifth project features a Tkinter GUI application that lets users filter a movie dataset by minimum release year and rating using Python libraries Tkinter, Pandas, and PandasTable. The filter_data() function updates the displayed table based on user inputs, while the refresh_data() function resets it to show the original dataset. The GUI includes fields for entering minimum release year and rating, buttons for filtering and refreshing, and a PandasTable for displaying the data. The application allows for interactive data filtering and visualization, with the table initially populated with sample movie data. In the sixth project, a retail store manager uses a DataFrame containing sales data to identify products that are both popular and profitable. By applying logical operators to filter the DataFrame, the goal is to isolate products that have sold more than 100 units and generated revenue exceeding $5000. This filtering is achieved using the Pandas library in Python, where the & operator combines conditions to select the relevant rows. The resulting DataFrame, which includes only products meeting both criteria, provides insights for decision-making and analysis in retail management. The seventh project involves creating a Tkinter-based GUI application to manage and visualize sales data. The GUI displays data in a table and a bar graph, allowing users to filter products based on minimum quantity sold and total revenue. The application uses pandas for data manipulation, pandastable for table display, and matplotlib for the bar graph. The GUI consists of an input frame for user filters and a display frame for showing the table and graph side by side. Users can update the table and graph by clicking "Filter Data" or reset them to the original data with the "Refresh" button, providing an interactive way to analyze sales performance. In chapter three, the first project demonstrates how to sort synthetic financial data for analysis. The code imports libraries, sets random seeds for reproducibility, and generates data for businesses including revenue and expenses. It then creates a DataFrame with this data, sorts it by monthly revenue in descending order, and saves the sorted DataFrame to an Excel file. This process aids in organizing and analyzing financial data, making it easier to identify top-performing businesses. The second project creates a Tkinter GUI to view and interact with synthetic financial data, displaying monthly revenue and expenses for various businesses. It generates random data, stores it in a DataFrame, and sets up a GUI with two tabs: one for sorting by revenue and another for expenses. Each tab features a table to display the data and a matplotlib plot for visual representation. The GUI allows users to sort and view data dynamically, with alternating row colors for readability and embedded plots for better analysis. The third project generates synthetic unemployment data for 10 regions over 5 years, sets random seeds for reproducibility, and creates a DataFrame with the data. It then sorts the DataFrame alphabetically by region and saves it to an Excel file named "synthetic_unemployment_data.xlsx". Finally, the script prints a confirmation message indicating that the data has been successfully saved. The fourth project generates synthetic unemployment data for 25 regions over a 5-year period and creates a Tkinter GUI for interactive data exploration. The data, organized into a DataFrame and saved to an Excel file, is displayed in a tabbed interface with two views: one sorted by unemployment rate and another by year. Each tab features scrollable tables and corresponding bar charts for visual analysis. The UnemploymentDataGUI class manages the interface, updating tables and graphs dynamically to allow users to explore regional and yearly unemployment variations effectively. The fifth project demonstrates how to concatenate dataframes with synthetic temperature data for various countries. Initially, we generate temperature data for countries like the USA and Canada for each month. Next, we create an additional dataframe with temperature data for other countries such as the UK and Germany. We then concatenate the original and additional dataframes into a single dataframe and save the combined data to an Excel file named combined_temperatures.xlsx. The steps involve generating synthetic data, creating additional dataframes, concatenating them, and exporting the result to Excel. The sixth project demonstrates how to build a Tkinter application to visualize synthetic temperature data. The app features a tabbed interface with tabs for displaying raw data, temperature graphs, and filters. It uses alternating row colors for better readability and includes functionality for filtering data by country and month. Users can view and analyze temperature data across different countries through tables and graphical representations, and apply or reset filters as needed. The seventh project demonstrates how to perform an inner join on two synthetic dataframes: one containing housing details and the other containing owner information. First, synthetic data is generated for houses and their owners. The dataframes are then merged on the common key, HouseID, using an inner join to include only rows with matching keys. Finally, the combined data is saved to an Excel file named combined_housing_data.xlsx. The result is an Excel file that contains details about houses along with their respective owners. The eight project provides an interactive platform for managing and visualizing synthetic housing data. Users can view comprehensive tables, apply filters for location and house type, and analyze house price distributions with Matplotlib plots. The application includes tabs for displaying data, filtering results, and generating visualizations, with functionalities to reset filters, save filtered data to Excel, and ensure a user-friendly experience with alternating row colors in tables and dynamic updates. To demonstrate an outer join on DataFrames with synthetic medical data, in ninth project, we create two DataFrames: one for patient information and another for medical records. We then perform an outer join to ensure all patients and records are included, even if some records don't have corresponding patient data. The code generates synthetic data, performs the outer join using pd.merge() on the PatientID column, and saves the result to an Excel file named outer_join_medical_data.xlsx. This approach provides a comprehensive dataset with complete patient and medical record information. The tenth project involves creating a Tkinter-based desktop application to visualize and interact with synthetic medical data. The application uses an outer join to merge patient and medical record datasets, displaying the comprehensive result in a user-friendly table. Users can filter data by patient ID and condition, view distribution graphs of medical conditions, and save filtered results to an Excel file. The GUI, leveraging Tkinter and Matplotlib, includes tabs for data display, filtering, and graph visualization, providing a robust tool for exploring medical datasets. In chapther four, the first project demonstrates creating and manipulating a synthetic insurance dataset. Using numpy and pandas, the script generates random data including columns for Policyholder, Age, State, Coverage_Type, and Premium. It groups this data by State and Coverage_Type to show basic data segmentation, then saves the dataset to an Excel file for further analysis. The code provides a practical framework for simulating and analyzing insurance data by illustrating the process of data creation, grouping, and storage. The second project demonstrates a Tkinter GUI application designed for analyzing a synthetic insurance dataset. The GUI displays 1,000 records of policyholder data in a scrollable table using the Treeview widget, with options to filter by state and coverage type. Users can save filtered data to an Excel file and generate a bar plot of policy distribution by state, integrated into the Tkinter window using Matplotlib. This application provides interactive tools for data exploration, filtering, exporting, and visualization in a user-friendly interface. The third project focuses on creating, analyzing, and aggregating a large synthetic sales dataset with 10,000 records. This dataset includes salespersons, regions, products, sales amounts, and timestamps, simulating a detailed sales environment. The core task involves grouping the data by region, product, and salesperson to calculate total sales and transaction counts. This aggregated data is saved to an Excel file, providing insights into sales performance and trends, which helps businesses optimize their sales strategies and make informed decisions. The fourth project develops a Tkinter GUI for analyzing synthetic sales data, allowing users to explore raw and aggregated data interactively. The application includes a dual-view setup with raw and aggregated data tables, filtering options for region, product, and salesperson, and visualization features for generating plots. Users can apply filters, view data summaries, save results to Excel, and visualize sales trends by region. The GUI is designed to provide a comprehensive tool for data analysis, visualization, and reporting. The dataset includes 10,000 records with attributes such as salesperson, region, product, sales amount, and date, and is grouped by region, product, and salesperson to aggregate sales data. The fifth project demonstrates how to create and analyze a synthetic transportation dataset. The code generates a large dataset simulating vehicle and route data, including distances traveled and durations. It groups the data by vehicle and route, calculating total and average distances and durations, and then saves these aggregated results to an Excel file. This approach allows for detailed examination of transportation patterns and performance metrics, facilitating reporting and decision-making. The sixth project outlines a Tkinter GUI project for analyzing synthetic transportation data using Python. This GUI, combining Tkinter and Matplotlib, provides a user-friendly interface to inspect and visualize large datasets involving vehicle routes, distances, and durations. It features interactive tables for raw and aggregated data, filter options for vehicle, route, and date, and integrates various plots like histograms and bar charts for data visualization. Users can apply filters, view dynamic updates, and save filtered data to Excel. The goal is to facilitate comprehensive data analysis and enhance decision-making through an intuitive, interactive tool. In chapter five, the first project involves generating and analyzing a synthetic dataset representing gold production across countries, years, and regions. The dataset, created with attributes like country, year, region, and production quantities, simulates complex real-world data for detailed analysis. By using the pivot_table method, the data is transformed to aggregate gold production metrics by country and region over different years, revealing trends and patterns. The results are saved as both original and pivoted datasets in Excel files for easy access and further analysis, aiding in decision-making related to mining and resource management. The second project creates an interactive Tkinter GUI to visualize and interact with a large synthetic dataset on gold production, including details on countries, regions, mines, and yearly production. Using pandas and numpy to generate the dataset, the GUI features multiple tabs for viewing the original data, pivoted data, and various summary statistics, alongside graphical visualizations of gold production trends across countries, regions, and years. The application integrates matplotlib for embedding charts within the Tkinter interface, making it a comprehensive tool for exploring and analyzing the data effectively. The third project demonstrates how to create a synthetic dataset simulating stock prices for multiple companies over 10,000 days, using random number generation to simulate stock prices for AAPL, GOOG, AMZN, MSFT, TSLA, and META. The dataset, initially in a wide format with separate columns for each company's stock prices, is then reshaped to a long format using pd.melt(). This long format, where each row represents a single date, stock, and its price, is often better suited for data analysis and visualization. Finally, both the original and unpivoted DataFrames are saved to separate Excel files for further use. The fourth project involves developing a visually engaging Tkinter GUI to analyze and visualize a synthetic stock dataset. The application handles stock price data for multiple companies, offering users both the original and unpivoted DataFrames, along with summary statistics and graphical representations. The GUI includes tabs for viewing raw and transformed data, statistical summaries, and interactive graphs, utilizing Tkinter's advanced widgets for a polished user experience. Data is saved to Excel files, and Matplotlib charts are integrated for clear data visualization, making the tool useful for both casual and advanced analysis of stock market trends. In chapter six, the first project demonstrates creating a large synthetic road traffic dataset with 10,000 rows using randomization techniques. Fields include Date, Time, Location, Vehicle_Count, Average_Speed, and Incident. Random NaN values are introduced into 10% of the dataset to simulate missing data. The dataset is then cleaned by removing rows with any missing values using dropna(), and the resulting cleaned DataFrame is saved to 'cleaned_large_road_traffic_data.xlsx' for further analysis. The second project creates a Tkinter-based GUI to analyze and visualize a synthetic road traffic dataset. It generates a dataset with 10,000 rows, including fields like date, time, location, vehicle count, average speed, and incidents. Random missing values are introduced and then removed by dropping rows with any NaNs. The GUI features four tabs: one for the original dataset, one for the cleaned dataset, one for summary statistics, and one for distribution graphs. Users can explore data tables with Tkinter's Treeview widget and view visualizations such as histograms and bar charts using Matplotlib, providing a comprehensive tool for data analysis. The third project generates a large synthetic electricity dataset to simulate real-world patterns in electricity consumption, temperature, and pricing. Missing values are introduced and then handled by filling gaps with regional averages for consumption, forward-filling temperature data, and using overall means for pricing. The cleaned dataset is saved to an Excel file, offering a valuable resource for testing data processing methods and developing data analysis algorithms in a controlled environment. The fourth project demonstrates a Tkinter GUI for handling missing data in a synthetic electricity dataset. The application offers a multi-tab interface to analyze electricity consumption data, including features for displaying the original and cleaned DataFrames, summary statistics, distribution graphs, and time-series plots. Users can view raw and processed data, explore statistical summaries, and visualize distributions and trends in electricity consumption, temperature, and pricing over time. The GUI integrates data generation, cleaning, and visualization techniques, providing a comprehensive tool for electricity data analysis.