Compute Percentiles for Parquet Files Online
Compute percentiles, median, MAD, mode, and kurtosis for numeric columns in Parquet files directly in your browser. Optionally group by a category column. Results download as CSV — no upload required.
About this tool
Compute distribution statistics for any numeric column in a Parquet file: median (p50), p25, p75, p90, p95, p99, MAD (Median Absolute Deviation), mode, kurtosis, and any custom percentile you specify. Optionally group by a category column to get per-group statistics. The output is a compact summary table ready to download as CSV. This goes beyond the basic sum/avg/min/max of the Aggregate tool — use it when you need to understand the full distribution of a column, spot skew, or calculate SLO targets. Runs locally in your browser.
Frequently Asked Questions
What is the difference between quantile_cont and median in a Parquet file?
median() is exactly equivalent to quantile_cont(col, 0.5) — the 50th percentile using linear interpolation. quantile_cont always interpolates between real values if the exact percentile falls between two data points. quantile_disc (discrete) always returns an actual value from the dataset.
What is MAD and when should I use it for Parquet data?
MAD stands for Median Absolute Deviation — it measures the spread of the data around the median. It is more robust than standard deviation because it is not affected by extreme outliers. Use it when your data has heavy tails or outliers that would inflate the standard deviation.
Can I compute percentiles per group in a Parquet file?
Yes. Set the Group By column and all selected statistics are computed independently within each group. The output has one row per group with all the requested statistics as columns.
Is my data private?
Yes — completely. Your file is never uploaded to any server. Everything runs locally in your browser using WebAssembly — processing happens entirely inside your tab. Once you close the tab, nothing is retained.
What is the maximum file size?
The free limit is 50 MB. For larger files, performance depends on your device's available memory — most modern machines handle 500 MB to 1 GB comfortably.
Related Tools
Compute Correlation Matrix for Parquet Files Online
Compute a Pearson correlation matrix for numeric columns in Parquet files directly in your browser. Instantly spot which variables move together — colour-coded heatmap, no upload required.
Find Fuzzy Duplicates in Parquet Files Online
Find near-duplicate rows in Parquet files using Levenshtein edit distance or Jaro-Winkler similarity — all in your browser. Set your own threshold and download the matched pairs as CSV — no upload required.
Hash & Anonymise Columns in Parquet Files Online
Anonymise or pseudonymise columns in Parquet files by replacing values with MD5, SHA-256, or DuckDB hashes — directly in your browser. Useful for GDPR compliance and sharing data without exposing PII — no upload required.
Group Concatenate Parquet Files Online
Concatenate text values within each group in Parquet files directly in your browser. Group rows by a column and join the values of another column into a single delimited string — no upload required.
Truncate Dates in Parquet Files Online
Truncate date and timestamp columns in Parquet files to a chosen precision — year, quarter, month, week, day, hour, or minute — directly in your browser. Rounds timestamps down to the start of each period. No upload required.