Skip to main content
Version: v0.7.1

Data Preprocessing

The Data Preprocessing page is used to quickly clean, normalize, and standardize whole-rock major element data before subsequent calculations (such as CIPW, plotting, and geochemical indices).

plot_home_ui

Module Features​

This module mainly covers common geochemical preprocessing tasks:

  • Import tabular data and recognize oxide columns
  • Normalize major elements on an anhydrous basis
  • Estimate iron valence distribution (FeO / Fe2O3)
  • Calculate common geochemical indices (such as Mg#, A/CNK, A/NK)
  • Export processed results for subsequent workflows

  1. Prepare data table
    Recommended to include standard oxide headers, such as SiO2, TiO2, Al2O3, FeOT, FeO, Fe2O3, MgO, CaO, Na2O, K2O, etc.

  2. Load data
    You can paste data directly into the table, or load sample data first to verify the workflow.

  3. Configure preprocessing options
    Enable or disable modules as needed:

    • Anhydrous normalization
    • Iron valence estimation method
    • Geochemical index calculation

  4. Run processing
    Run preprocessing and review the output worksheet content.

  5. Export results
    Export results as .csv for subsequent analysis or plotting.

Key Parameter Descriptions​

1) Anhydrous Normalization​

When enabled, volatile components (such as LOI, H2O, CO2) are deducted and major oxides are rescaled to 100%.

Enable this option when you need cross-sample comparable major element analysis.

2) Iron Valence Estimation​

When data only has total iron (FeOT), FeO and Fe2O3 distribution can be estimated according to strategy:

  • Auto-estimate from total iron
  • Back-calculate based on existing FeO / Fe2O3
  • Correct using empirical ratio Fe3+/Fe

Please choose the appropriate method based on data quality and research subject.

3) Geochemical Indices​

The module supports calculating:

  • Mg#
  • A/CNK
  • A/NK

These indices are commonly used for magma evolution and aluminum saturation interpretation.

Input & Output​

Input​

  • Worksheet containing sample IDs and oxide columns
  • Recommended to use wt.% numeric format

Output​

  • Cleaned/normalized columns (depending on enabled modules)
  • Iron valence estimation columns (when enabled)
  • Geochemical index columns (when enabled)
  • Processing summary information (for tracking parameters and results)

Usage Recommendations​

  • It is recommended to keep original data in a separate worksheet to avoid overwriting.
  • Confirm column name conventions before processing to reduce field recognition errors.
  • When used for papers or reports, it is recommended to clearly record the selected preprocessing parameters in the methods section.

If you would like to add new preprocessing strategies, please submit an issue or discussion in the project repository.