Mineral-exploration data workbench

Read the rock.
Cut the noise.

Raw geochemistry arrives dirty: detection-limit spikes, unit mismatches, fat outlier tails. Oxide Explorer is the bench where you load a provincial dataset, read every element-oxide relationship at once, and cut the noise by hand or by recipe until the signal stands clear.

Open the Explorer Browse the catalog
The workbench

Five areas, one dataset

Each area does one job well. Hover any card for what it is; open it when you need it.

01
Clean

Explorer

Load a geochem CSV and read every element-to-oxide pair as a log-log scatter. Draw a polygon, fit a line, strip the outliers, clean each distribution bin by bin.

Open
02
Add

Upload

Bring your own geochemistry. Drop a CSV and it lands in cloud storage, ready to open in the Explorer next to the built-in datasets.

Open
03
Browse

Catalog

Browse the open geoscience datasets in cloud storage. Weigh size, columns, and coverage before you commit one to a scan.

Open
04
Rank

Targets

Score a region for prospectivity. Pick a commodity and a box on the map; the model ranks where the next deposit likely hides.

Open
05
Automate

Pipelines

Turn one cleaning session into a recipe. Replay the same ordered steps across every dataset in a single hands-off run.

Open
06
Extend

Custom tools

Write a cleaning operation as an expression. Validate it on real rows, then save it as a versioned step any recipe can use.

Open
The flow

From a raw CSV to a ranked map

  1. 01
    Pick a dataset
    from the Catalog
  2. 02
    Read the pairs
    element vs oxide
  3. 03
    Clean the noise
    bins and outliers
  4. 04
    Save a recipe
    replay it anywhere
  5. 05
    Rank the ground
    prospectivity map