Different inputs.
One place to explore.
Numeric
CSV input, dataset overview, statistical summaries, distributions, and relationships.
Text
Direct input and file-based text, sentiment, topics, word clouds, and frequencies.
Images
Image uploads, color and property analysis, model prediction, and custom training controls.
Real-time
Stock-history and news-analysis paths connected to external data services.
Feature map based on the repository’s analysis-type menu and page functions.
Start with the table.
Look beyond the totals.
The numeric page reads CSV data with pandas and plots numeric columns with Plotly. Its overview presents rows, columns, missing-value counts, a preview, and data types.
A separate branch looks for a final column named with “target,” “prediction,” or “class,” then requests class balance, correlations, feature importance, and feature distributions from the numeric-analysis module.
The numeric uploader in app.py accepts CSV. The README also mentions Excel, but the reviewed numeric-page control does not expose Excel uploads.
Words and pixels
need different questions.
| Input | Interface workflow |
|---|---|
| Text | Direct entry or TXT, CSV, PDF, and DOCX upload controls; statistics, sentiment, topics, word clouds, and word frequencies. |
| Images | PNG, JPG, and JPEG uploads; property and color analysis; a custom-model prediction path. |
| Training data | A local directory organized into class folders, with training controls and progress displays. |
The custom-image path attempts to load best_model.pth. The repository includes yolov8n.pt; it does not publish a custom training dataset or evaluation scores. Model weights and interface controls alone do not establish prediction accuracy.
A Python workspace
with adjustable views.
The external-data page separates stock and news workflows. It requests historical prices, technical-analysis outputs, forecasts, news articles, and sentiment results through the real-time analysis module.
This case study documents the public interface code. The imported modules.numeric_analysis, modules.text_analysis, modules.image_analysis, and modules.realtime_analysis files are absent from the reviewed repository snapshot, so the app cannot run from that snapshot unchanged. No runtime results or model benchmarks are claimed here. Review the application code.