Column Extractor

WHAT IT DOES

Column Extractor is a browser extension that analyzes tables on any webpage and identifies which columns contain numeric data. It displays a structured list showing each column's name, data type, value range (minimum and maximum), average, and count of numeric cells. For each numeric column, you can view sample values and see the percentage of rows that contain numeric data versus text. The tool lets you copy the extracted numeric column data in two formats: CSV for spreadsheets or JSON for programmatic use. This is ideal for data analysts, researchers, and anyone working with tabular data online who needs to quickly isolate and export numerical information without manual copying.

WHERE IT RUNS

Column Extractor works on any website that displays data in HTML tables. It runs entirely in your browser as a lightweight userscript, detecting table structures on any page you visit. It works on financial dashboards, scientific data portals, e-commerce product lists with pricing, sports statistics pages, research databases, inventory systems, and internal business tools. The tool handles static tables, dynamically updated tables, and even tables inside iframes where permitted. All data processing occurs locally—no information leaves your browser.

HOW TO USE

After installation, open the floating panel by clicking the extension icon or using the keyboard shortcut. The panel automatically scans the current page and lists every table found. Use the dropdown menu to select a table, then choose a specific column from that table. The panel immediately displays detailed statistics for that column: the number of numeric values, the range of values, the average, and sample entries. To copy the data, simply click the CSV button to copy comma-separated values to your clipboard, or click the JSON button to copy a structured JSON object. The panel remembers your last selected table and column per website, so you can return to your analysis quickly.

REAL MECHANISM — WHY THIS WORKS

Most data extraction tools attempt to parse table content as a flat text stream, which fails when tables contain merged cells, headers, footers, or mixed data types. Column Extractor uses a different approach: it traverses the DOM tree and maps each table cell to its logical row and column position, accounting for rowspan and colspan attributes. It then applies a type inference engine that examines the content of each cell through multiple passes. The first pass collects raw text content and stores it with its cell coordinates. The second pass normalizes the text by removing non-numeric characters while preserving decimal separators and negative signs. The third pass applies a set of pattern matchers that distinguish true numbers from phone numbers, dates, postal codes, and product identifiers using context-aware heuristics. Once a column is identified as numeric, the engine computes statistical aggregates directly from the parsed values without converting the DOM to a string or using regular expressions that can be fooled by localization. This ensures that the extraction is both accurate and resilient across thousands of different table structures.

THE PANEL

The floating panel is fully draggable and remembers its position per site, snapping to the edges of the viewport when dragged near them. You can collapse the panel to a small icon to minimize screen distraction and expand it back with a single click. The panel is resizable from the bottom-right corner, giving you more space to view column statistics or sample data. Press the Escape key to close the panel instantly. The panel remains open across page navigation within the same domain, maintaining your current selection and analysis results.

PLEASE NOTE

Column Extractor requires the free BotGentz extension to be installed and enabled, as it relies on the cross-extension messaging API provided by BotGentz for table detection and DOM access permissions. Because websites frequently update their structure, table detection and column analysis may require periodic script updates to maintain compatibility. If a table is not detected correctly, please check for updates or report the issue. All data processing occurs entirely in your browser. No table data, column values, or statistical results are ever sent to any server. The extension does not read, store, or transmit any personal information. It only reads the DOM of the pages you choose to interact with and writes copied data to your clipboard at your explicit request.