RunTangent
DATA & ANALYTICS / CSV INSIGHTS

CSV Insights

Explore CSV data with summaries, charts, filters and pivot tables.

Find the story inside your rows.

No file loaded

A place for patterns to come into view.

Load a CSV to explore distributions, compare groups, and inspect your records. You can also drop your file here.

Rows: 0 Columns: 0
Row filters

Rules combine with AND and your search slice. Numeric/date bounds are inclusive; missing or invalid values do not match ranges. Date bounds use YYYY-MM-DD; source slash dates use month/day/year.

These controls change the cards shown, not the dataset or analysis slice.

Inspect the current slice. A row export includes all matching records across pages.
Showing rows 0 - 0 of 0
FROM RECORDS TO RELATIONSHIPS

A clearer view.
A better question.

Explore CSV data in your browser, from column distributions to grouped summaries. Start with the overview, then investigate the details behind it.

How to explore a dataset

  1. Open your CSV. Load a file and review the detected columns in Auto-Profiler. Check the distributions and summary statistics before drawing conclusions.
  2. Ask a focused question. Switch to Manual Explorer, choose a grouping column and measure, and select the chart and aggregation that suit your question.
  3. Check the underlying rows. Use Row filters for exact categories or inclusive numeric/date bounds, combined with Slice Data search. Rules use AND. Inspect matching records in Raw Data Grid; clear row rules and search separately to restore the full dataset.
  4. Keep your findings. Save a chart as PNG, export grouped pivots from Manual Explorer, or export all matching grid records across pages. A chart drill-down narrows the grid further; clearing it returns to the current slice. Download results before closing or replacing the dataset.

Understand the columns

Review distributions and numeric summaries, including quartiles and the 90th percentile. Find columns by name or type; these card controls do not filter dataset rows. Open Understand these statistics for calculation and exclusion details.

Compare your groups

Choose dimensions and measures in Manual Explorer, build a chart and download the grouped result.

Inspect the records

Use Raw Data Grid to check the underlying rows. Keep your original file; this workspace is not an automatic save.

A FEW THINGS WORTH KNOWING

Explore with
confidence.

What the views show, and how to keep your results.

Back to the workspace ↑
What can I explore in CSV Insights?

Load a CSV to profile columns, build grouped charts in Manual Explorer, or inspect records in Raw Data Grid. Search profile cards by column name or type without filtering rows. This tool explores data; it does not edit or clean source values.

Does my dataset leave my browser?

CSV parsing and analysis run in this page and its background worker. No dataset upload is needed. Work is held in browser memory, not automatically saved. Download results before closing, reloading or replacing the dataset.

How do row filters, search and chart drill-downs work together?

Row rules combine with AND and the Slice Data search. Categories match exact text and case; numeric/date bounds are inclusive and exclude missing or invalid values. Date bounds use YYYY-MM-DD; source slash dates use month/day/year. Clear rules and search separately. Chart drill-downs narrow the grid further; clearing a drill-down returns to the underlying slice.

What can I download?

Save charts as PNG, export grouped pivot results from Manual Explorer as CSV, or export all records matching the current Raw Data Grid view across pages. Row exports preserve source values; pivot exports contain calculated groups. A grid drill-down does not change the underlying Manual Explorer pivot.

How should I interpret the statistics and limits?

Numeric profiles exclude blanks and nonnumeric values. Quartiles and the 90th percentile use linear interpolation; standard deviation uses the population convention. IQR outliers are review flags, remain included in statistics and are not automatically removed. Device memory limits apply. Automatic types and charts support exploration, not statistical guarantees; use Correlation Matrix for dedicated relationship analysis.