On this page
Publish Parquet and GeoParquet Datasets
Parquet is a column-oriented data format suited to analytical queries. GeoParquet adds geometry information for spatial datasets. MiniUp lets you convert supported files into hosted datasets for dashboards, maps, and other apps.
Publish a dataset
- Open Dashboard → your Site → Parquet.
- Under Upload data, choose a source file: CSV, XLSX, XLS, ZIP shapefile, GeoJSON, JSON, Parquet, or GeoParquet.
- Set the dataset name and slug, and review the detected schema and sample rows.
- For spatial data, verify the geometry or select the correct latitude and longitude fields. If detection is ambiguous, choose both columns explicitly.
- Complete the conversion/publish action and wait for a successful result.
- Open the published dataset’s detail view, inspect its format, row count, size, and access.
- Use Use this data to copy the relevant app, map, or data usage instructions.
Choose Parquet instead of a Table
Use Parquet for analytical reads and larger data files that do not need individual record editing. Use a MiniUp Table or editable layer when users must create or update records. Replacing or refreshing a dataset is different from editing a single Table record.
Keep large-data apps responsive
Select only the columns needed, aggregate before charting, and page through rows. Avoid turning the entire dataset into JavaScript objects just to display a small table. Browser memory still matters; a compact file can require more memory after loading and processing.
Example: a regional sales dashboard
Upload a sales CSV, inspect the exact region and amount field names, publish Parquet, then use DuckDB in the browser to calculate regional totals and display one page of records at a time.