How morisde/financial-statements works
Main Components
• Data Ingestion – This component represents the source of the dataset: the “Financial Statements.csv”.
• Data Analysis/EDA – This module covers both the generation of the in-depth EDA report (via ydata_profiling) and the visual analytics (including SNS.pairplot explorations).
• Machine Learning Pipeline – The section where PyCaret is employed to set up the regression model (using Lasso) to predict Gross Profit.
• Output/Reporting – This includes the final visualization results in the Jupyter Notebook and the generated “report.html”.
Structure and Relationships
• Data Input (Financial Statements.csv): At the beginning of the pipeline, the raw CSV data is imported into the Jupyter Notebook (“Financial_Statements_of_Major_Companies(2009_2023).ipynb”).
• EDA Process: From the raw data, an exploratory step is performed. Use the ydata_profiling library to build an in-depth report, and include preliminary visualizations like SNS.pairplot (noting that this function may be optionally skipped due to execution time). This module serves two purposes:
• Data understanding and validation.
• Informing any subsequent cleaning or preprocessing steps.
• Machine Learning (ML) Execution: Following EDA, the PyCaret library is used. This part of the notebook is responsible for:
• Preparing a ML model (Lasso regression) for predicting Gross Profit.
• Evaluating model performance by comparing predicted vs. actual values in the final visualization.
• Reporting/Visualization Output: The final results – including the execution of predictive modeling and final comparison plots – are saved and/or rendered as “report.html” and may also be restored in the notebook.