Dynamic Graph Generation from Excel Using Machine Learning Algorithm Data Visualization Dashboard

Overview

This research describes a systematic pipeline for automated conversion of Excel spreadsheets into publication-quality dynamic visualizations through integrated machine learning, automated schema detection, and interactive dashboard functionality. The system addresses the gap between raw tabular data and effective visual communication by combining algorithmic chart recommendation, parameterization, and interactive exploration capabilities. The approach encompasses the complete workflow from data ingestion through vector graphic export.

Methods and approach

The pipeline incorporates multiple sequential components: dataset ingestion mechanisms for Excel format data, automatic schema detection to identify data types and structures, feature engineering to extract visualization-relevant properties, machine learning-based chart type recommendation with parameter optimization, graph rendering functionality, and an interactive web-based dashboard enabling exploration and export operations. The system architecture integrates anomaly detection and automated labeling mechanisms to enhance visualization informativeness. Validation employed three real-world datasets representing distinct domains: financial data, sensor time-series measurements, and survey response collections.

Key Findings

Quantitative and qualitative evaluation demonstrated improvements across multiple dimensions compared to manual visualization selection. The approach yielded measurable reductions in time-to-visualization metrics and enhanced user satisfaction scores. Automated chart recommendation and parameterization demonstrated increased accuracy in selecting appropriate visualization types relative to manual selection processes. The system successfully generated exportable vector graphics with integrated anomaly highlighting and automated labeling.

Implications

The presented methodology establishes a reproducible framework for reducing manual effort in data visualization workflows while maintaining publication-quality output standards. The integration of machine learning-driven recommendations addresses the technical barriers associated with chart type selection and parameterization, potentially reducing cognitive load for practitioners operating within constrained time environments. The dashboard component extends beyond visualization generation to enable exploratory analysis and output customization.

Scope and limitations

This summary is based on the study abstract and available metadata. It does not include a full analysis of the complete paper, supplementary materials, or underlying datasets unless explicitly stated. Findings should be interpreted in the context of the original publication.