✦ PIXELOKAY · INTELLIGENT DATA TOOLS

Multidimensional Data Explorer

Compare dimensions, reveal patterns and filter large datasets—without sending your data anywhere.

Unsaved · lives in this tab

TOTAL ROWS0

FILTERED ROWS0

DIMENSIONS0

ACTIVE FILTERS0

Drag vertically on a parallel-coordinate axis to select a range. Use the controls panel for precise filters.

Dimensional insightsSample-aware

Preview selected dataFirst 12 rows

Exploratory visualizations are not statistical proof. Scatterplots draw a deterministic subset when necessary; displayed row counts and correlations use all filtered rows. Missing values are excluded from affected computations.

Private multidimensional data visualization online

Explore numerical relationships with parallel-coordinate graphs, scatterplot matrices, and Pearson correlation matrices. Import CSV, TSV, or flat JSON locally, brush numeric axes to select rows, filter categorical groups, and export visualizations or filtered data without uploading private datasets.

How comparisons work

Rows with missing numeric values are omitted from plots requiring those coordinates. Pearson correlations use every complete pair in the selected dataset; correlations for constant columns are undefined. For responsiveness, the parallel and scatter chart views draw representative samples, while the filter statistics count every selected row.

Using Multidimensional Data Explorer

Explore CSV/TSV data with filters, parallel coordinates, scatter comparisons, and Pearson correlations, then export charts, filtered rows, reports, or backups.

Workflow

Import a CSV/TSV table and confirm the detected columns while keeping original cell values available for review.

Filter and group rows, choose up to six numeric axes, and inspect parallel coordinates, scatter comparisons, and summary statistics.

Review which visuals are sampled and which statistics use the full filtered rows, then export the chart, data, report, or project backup.

Worked example

Import a table of measurements with a category and several numeric columns. Filter to one category, compare two numeric fields in a scatter view, and check the Pearson correlation before exporting the filtered rows and chart.

Limits and data handling

Visual points are sampled, while counts and statistics use all filtered rows. Pearson correlation does not establish significance or causation. Original cell values remain separate from inferred plotting values; missing values are not imputed. Limits are 40,000 rows, 40 columns, 8 MB input, 20 MB backup, and six numeric axes. Optional browser copies are not durable backups.

Frequently asked questions

Do the charts plot every filtered row?

Visualizations sample points for display, while counts and statistics use the complete filtered population.

Does a Pearson correlation prove that two fields are related causally?

No. It summarizes a linear association in the selected data and does not establish significance or causation.