This page was generated from docs/examples/mineralML_mapping.ipynb. Interactive online version: .

Python Notebook Download

""" Created on November 13, 2023 // Updated on March 20, 2026 // @author: Sarah Shi """

import os
import numpy as np
import pandas as pd

import mineralML as mm

import matplotlib.pyplot as plt
%matplotlib inline
%config InlineBackend.figure_format = 'png'

mineralML Quickstart for Mapped EDS Data

This notebook shows how to load and run your quantitative EDS data through mineralML with an example gabbroic nodule from the Galapagos: 09g3. These data are from Gleeson et al., 2024 (find the paper here: https://doi.org/10.1093/petrology/egaf031) in this GitHub repository (https://github.com/gleesonm1/GleesonEtAl_JPet_2024_supplement/tree/main/Code_Figures/Data/LargeScaleMaps/Gabbro%20Samples). Please refer to the paper for more information about this sample.

This is a five step process:

1. Load a directory containing all your CSVs of mapped chemical data with mm.load_df (or pd.read_csv directly). [Optional] Convert the input from element to oxide wt%.

2. Predict the mineral class with mineralML and automatically plot the mineral phase map, mineral phase counts, and prediction score histograms

3. Plot prediction score map and individual oxide maps.

4. Plot compositions of mapped minerals in various classification diagrams (ternary, quadrilateral).

5. Plot chemical variation maps.

I have conveniently (I hope!) wrapped all of these bits into one function, called mm.run_map.

We loaded in the mineralML Python package as mm. mineralML has trained machine learning models for classifying minerals. This implementation aims to get your electron microprobe or quantitative EDS compositions classified and processed. We remove some degrees of freedom to simplify the process as much as possible. The minerals considered for this study include: Amphibole, Apatite, Biotite, Calcite, Chlorite, Epidote, Feldspar (Alkali Feldspar and Plagioclase), Garnet, Glass, Kalsilite, Leucite, Melilite, Muscovite, Nepheline, Olivine, Oxide (Rhombohedral_Oxides including Hematite-Ilmenite, Spinel_Group including Magnetite-Spinel), Pyroxene (Clinopyroxene, Orthopyroxene, Na-Pyroxene), Quartz, Rutile, Serpentine, Titanite, Tourmaline, and Zircon.

CSV files containing your mapped data in oxide weight percentages (or converted to) is necessary. Find an example here, reproduced from the Gleeson et al., 2024 GitHub linked above. The necessary oxides are SiO\(_2\), TiO\(_2\), Al\(_2\)O\(_3\), FeO\(_t\), MnO, MgO, CaO, Na\(_2\)O, K\(_2\)O, Cr\(_2\)O\(_3\), P\(_2\)O\(_5\), and ZrO\(_2\) (if you are aiming to classify zircon). For the oxides not analyzed for specific minerals, the preprocessing will fill in the nan values as 0.

1. Load and prepare data for analysis

Here, we will work with data that are in elemental weight percent. This means that we will have to do a conversion to oxide weight percent. The data directory is Maps/09g3.

# Find your directory of mapped mineral data, stored in Maps/09g3.
# This code identifies any file with CSV and appends it to the map.

base = "Maps"
map_dirs = []
for root, subdirs, files in os.walk(base):
    # Skip any path that includes 'Ignore' in its folder names
    if "Ignore" in root.split(os.sep):
        continue

    if any(f.lower().endswith(".csv") for f in files):
        map_dirs.append(root)

print(map_dirs)

['Maps/09g3', 'Maps/MountHood_MH0811b']

2. Apply the trained neural network with mm.run_map

We will use mm.run_map which will return all you need!

# Inspect the inputs and outputs of mm.run_map
help(mm.run_map)

Help on function run_map in module mineralML.mapping:

run_map(sample_input, renormalize=False, total_threshold=None, n_iterations=50, min_frac=1e-05, pred_score_threshold=0.6, units='element_wt%', top_k=None, phases=None, exclude_phases=None, phase_colors=None, bar_style='vertical', components_spec=None, remove_islands_flag=False, fill_holes_flag=False, hole_size=10, scalebar_um=None, pixel_size_um=None, scalebar_loc='lower left', scalebar_col='black', show=True)
    Load, convert, predict, and plot for one folder of CSV maps.
    Always computes mineral components and returns a full results dictionary.
    Use ``remove_islands_flag`` and ``fill_holes_flag`` to clean the phase map
    before plotting and downstream analysis.

    Parameters:
        sample_input (str | Path | dict): Directory path or a dict of oxide maps.
        renormalize (bool): If True, scale each pixel so oxides sum to 100 wt%.
            Applied after total masking.
        total_threshold (float | None): Pixels with oxide total below this
            value (wt%) are set to NaN before renormalization and prediction.
            Use this to mask epoxy/background.
        n_iterations (int): MC forward passes for prediction.
        min_frac (float): Minimum pixel fraction required to keep a phase.
        pred_score_threshold (float): Label NaN where max prediction score <
            threshold.
        units (str): Input format — 'element_wt%' or 'oxide_wt%'.
        top_k (int|None): Cap displayed phases after filtering.
        phases (list[str]|None): Explicit phases to plot (overrides auto-pick).
        exclude_phases (list[str]|None): Phases to remove from auto-pick.
        phase_colors (dict|None): Manual color mapping {PhaseName: HexColor}.
        bar_style (str): "vertical" for the default bar chart
            (``plot_phase_counts``), or "stacked" for a stacked horizontal
            bar (``plot_phase_proportions``).
        components_spec (dict|None): Custom mineral formula logic.
        remove_islands_flag (bool): If True, removes isolated pixel clusters
            smaller than 2 pixels (4-connected) from the phase map. Useful
            for cleaning up salt-and-pepper noise in the epoxy region.
        fill_holes_flag (bool): If True, fills enclosed background holes within
            continuous phase regions up to ``hole_size`` pixels. Useful for
            patching small gaps inside large mineral grains.
        hole_size (int): Maximum hole area (in pixels) to fill when
            ``fill_holes_flag`` is True.
        scalebar_um (float, optional): Length of the scale bar in micrometers.
        pixel_size_um (float): Physical size of a single pixel in micrometers.
        scalebar_loc (str): Location of the scale bar (e.g., 'lower left').
        scalebar_col (str): Color of the scale bar text/line.
        show (bool): If True, calls plt.show().

    Returns:
        result (dict): Dictionary with keys 'figs', 'shape', 'oxide_maps',
            'df_pred', 'mineral_map', 'pred_score_map', 'kept_phases',
            'component_maps', 'component_frames'. ``oxide_maps`` includes a
            ``'Total'`` key with the per-pixel oxide sum.

# Here is our all in one function! Read the inputs and outputs provided above.

output = mm.run_map(next((s for s in map_dirs if '09g3' in s), None), # provide the directory of interest. alternatively, you can provide the preloaded dictionary of oxides.
                    renormalize=False, # optionally renormalize totals to 100 wt%.
                    total_threshold=None, # optionally filter out SiO2 values below a given value, for when EDS picks up epoxy pixels
                    pred_score_threshold=0.6, # provide a prediction score threshold. here, i only want values with >= 0.6 prediction score.
                    min_frac=0.01, # provide a minimum pixel fraction for the phase to be displayed
                    units='element_wt%', # provide the unit. can choose 'element_wt%' or 'oxide_wt%'
                    phases=['Plagioclase', 'Clinopyroxene', 'Orthopyroxene', 'Oxide', 'Olivine', 'Glass'], # phases of interest
                    scalebar_um=50, # define size of scalebar desired, in microns
                    pixel_size_um=2, # define size of each pixel of scalebar, in microns
                    scalebar_loc='upper right', # specify location for scalebar
                    scalebar_col='white', # specify color for scalebar
                    )

[ok] Si Wt%Montaged Map Data.csv  →  Si  (329, 283)
[ok] Ca Wt%Montaged Map Data.csv  →  Ca  (329, 283)
[ok] Cr Wt%Montaged Map Data.csv  →  Cr  (329, 283)
[ok] Mg Wt%Montaged Map Data.csv  →  Mg  (329, 283)
[ok] Mn Wt%Montaged Map Data.csv  →  Mn  (329, 283)
[ok] Al Wt%Montaged Map Data.csv  →  Al  (329, 283)
[ok] Ti Wt%Montaged Map Data.csv  →  Ti  (329, 283)
[ok] Na Wt%Montaged Map Data.csv  →  Na  (329, 283)
[ok] Fe Wt%Montaged Map Data.csv  →  Fe  (329, 283)
[ok] K Wt%Montaged Map Data.csv  →  K  (329, 283)
mineralML: 93107 rows — 57725 classified by neural network, 7 by empirical rules (Zircon: 0, SiO2 polymorph: 0, Carbonate: 7), 35375 skipped (invalid/empty)
prep_df: 57725 row(s) processed (of 57725 input, 0 dropped).

/home/docs/checkouts/readthedocs.org/user_builds/mineralml/checkouts/stable/src/mineralML/hybrid.py:313: UserWarning: The following columns were missing and have been filled with NaN: ['Mineral']
  df = prep_df(df)

# Inspect what is in the outputs
output.keys()

dict_keys(['figs', 'shape', 'oxide_maps', 'df_pred', 'mineral_map', 'pred_score_map', 'kept_phases', 'component_maps', 'component_frames'])

Let’s say you now want to work with these data in dataframe form rather than dictionary form. How would you do this? Access output['df_pred'] to retrieve the predictions dataframe.

# Pull the dataframe of predictions
df_pred = output['df_pred']
display(df_pred)

	SiO2	TiO2	Al2O3	FeOt	MnO	MgO	CaO	Na2O	K2O	Cr2O3	P2O5	ZrO2	Predict_Mineral	Prediction_Score	Prediction_Score_Sigma	Second_Predict_Mineral	Second_Prediction_Score	Submineral
0	49.263735	0.246671	31.267656	-0.141366	-0.133547	-0.041928	11.384471	3.906129	0.287173	0.206763	NaN	NaN	Plagioclase	0.998627	0.005443	Muscovite	0.001223	Labradorite
1	52.781465	1.223476	29.213528	1.751936	-0.008407	0.353991	13.811718	3.613949	0.326185	0.009850	NaN	NaN	Plagioclase	0.999909	0.000345	Titanite	0.000041	Labradorite
2	49.582496	-0.666395	31.217623	0.669667	0.312345	0.135481	13.362425	2.559899	0.404478	-0.008271	NaN	NaN	Plagioclase	0.999078	0.002773	Muscovite	0.000342	Bytownite
3	49.766286	-0.061016	30.748715	-0.049472	0.052490	-0.063224	13.924789	3.722859	0.031619	-0.244754	NaN	NaN	Plagioclase	0.987317	0.083828	Leucite	0.007627	Labradorite
4	50.775667	0.283931	29.405951	0.363973	0.161280	0.434312	12.246722	2.495524	0.260420	0.922368	NaN	NaN	Plagioclase	0.997432	0.005905	Glass	0.001690	Bytownite
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
93102	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	NaN	NaN	None	NaN	NaN	NaN	NaN	NaN
93103	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	NaN	NaN	None	NaN	NaN	NaN	NaN	NaN
93104	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	NaN	NaN	None	NaN	NaN	NaN	NaN	NaN
93105	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	NaN	NaN	None	NaN	NaN	NaN	NaN	NaN
93106	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	None	NaN	NaN	NaN	NaN	NaN

93107 rows × 18 columns

3. Plot oxide concentration maps and prediction score maps

Let’s plot the original oxide maps loaded from the directory. We can examine how well the predicted phase map matches some of the observations made in oxide space. We have a handy function for doing so, with mm.plot_oxide_map.

fig, ax = mm.plot_oxide_map(
    output, # take the output from run_map
    oxide_name='SiO2', # specify the oxide of interest
    scalebar_um=50, # define size of scalebar desired, in microns
    pixel_size_um=2, # define size of each pixel of scalebar, in microns
    scalebar_loc='upper right', # specify location for scalebar
    scalebar_col='black', # specify color for scalebar
)

Let’s plot the prediction scores from the output, in mapped form. This allows for further investigation to determine where predictions are more and less certain.

fig, ax = mm.plot_score_map(
    output, # take the output from run_map
    scalebar_um=50, # define size of scalebar desired, in microns
    pixel_size_um=2, # define size of each pixel of scalebar, in microns
    scalebar_loc='upper right', # specify location for scalebar
    scalebar_col='black', # specify color for scalebar
)

4. Plot compositions of mapped minerals in various classification diagrams (ternary, quadrilateral).

We can do some more with mineralML now. Let’s plot all the feldspars, pyroxenes, and spinels in ternary space.

Identify the phases present.

# Here are all our feldspars
fspars = df_pred[df_pred.Predict_Mineral == 'Plagioclase']
display('Feldspars:', fspars)

# Here are all our pyroxenes
pxs_names = ['Clinopyroxene', 'Orthopyroxene']
pxs = df_pred[df_pred.Predict_Mineral.isin(pxs_names)]
display('Pyroxenes:', pxs)

# Here are all our oxides
ox_names = ['Oxide']
oxs = df_pred[df_pred.Predict_Mineral.isin(ox_names)]
display('Oxides:', oxs)

'Feldspars:'

	SiO2	TiO2	Al2O3	FeOt	MnO	MgO	CaO	Na2O	K2O	Cr2O3	P2O5	ZrO2	Predict_Mineral	Prediction_Score	Prediction_Score_Sigma	Second_Predict_Mineral	Second_Prediction_Score	Submineral
0	49.263735	0.246671	31.267656	-0.141366	-0.133547	-0.041928	11.384471	3.906129	0.287173	0.206763	NaN	NaN	Plagioclase	0.998627	0.005443	Muscovite	0.001223	Labradorite
1	52.781465	1.223476	29.213528	1.751936	-0.008407	0.353991	13.811718	3.613949	0.326185	0.009850	NaN	NaN	Plagioclase	0.999909	0.000345	Titanite	0.000041	Labradorite
2	49.582496	-0.666395	31.217623	0.669667	0.312345	0.135481	13.362425	2.559899	0.404478	-0.008271	NaN	NaN	Plagioclase	0.999078	0.002773	Muscovite	0.000342	Bytownite
3	49.766286	-0.061016	30.748715	-0.049472	0.052490	-0.063224	13.924789	3.722859	0.031619	-0.244754	NaN	NaN	Plagioclase	0.987317	0.083828	Leucite	0.007627	Labradorite
4	50.775667	0.283931	29.405951	0.363973	0.161280	0.434312	12.246722	2.495524	0.260420	0.922368	NaN	NaN	Plagioclase	0.997432	0.005905	Glass	0.001690	Bytownite
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
92980	50.751835	0.053460	30.892806	0.040494	-0.461759	0.181496	13.219259	4.255093	0.562793	0.125117	NaN	NaN	Plagioclase	0.998067	0.008614	Glass	0.001712	Labradorite
92981	47.055855	0.435257	28.540332	0.088852	-0.089822	0.027772	10.548025	3.540654	0.550766	0.032374	NaN	NaN	Plagioclase	0.980982	0.046478	Glass	0.012173	Labradorite
92982	49.572227	0.371584	32.248626	0.873017	-0.285326	0.130899	12.215584	4.380849	0.471066	0.514502	NaN	NaN	Plagioclase	0.999816	0.000423	Muscovite	0.000078	Labradorite
92983	49.374082	-0.392758	32.780453	0.389758	0.339789	0.275332	13.295722	4.352765	0.428836	0.269782	NaN	NaN	Plagioclase	0.997028	0.018393	Leucite	0.001778	Labradorite
92984	47.506399	0.326601	30.147186	0.261584	-0.122220	-0.125815	12.209264	4.127070	0.188657	-0.821505	NaN	NaN	Plagioclase	0.990222	0.040914	Glass	0.007537	Labradorite

32058 rows × 18 columns

'Pyroxenes:'

	SiO2	TiO2	Al2O3	FeOt	MnO	MgO	CaO	Na2O	K2O	Cr2O3	P2O5	ZrO2	Predict_Mineral	Prediction_Score	Prediction_Score_Sigma	Second_Predict_Mineral	Second_Prediction_Score	Submineral
49	46.703956	2.306829	8.091728	7.608155	0.179336	10.411849	18.795262	2.432519	0.532825	0.203853	NaN	NaN	Clinopyroxene	0.907966	0.130830	Glass	0.059600	Diopside
610	48.659501	1.695958	7.334367	7.926977	-0.333935	10.416438	17.438423	2.285797	1.183322	0.106932	NaN	NaN	Clinopyroxene	0.921416	0.161371	Glass	0.050355	Diopside
616	48.778790	1.189971	2.368374	9.717480	-0.178262	19.703388	16.359033	0.808957	0.195681	0.046046	NaN	NaN	Clinopyroxene	0.796175	0.219379	Amphibole	0.191745	Augite
626	43.817565	1.304287	4.129765	5.646358	-0.037404	11.766227	22.258578	1.406237	0.008721	0.322146	NaN	NaN	Clinopyroxene	0.974739	0.057883	Melilite	0.013015	Wollastonite
900	48.447279	1.397534	3.416259	9.859029	0.064421	16.668941	17.343585	1.357627	0.303317	0.192969	NaN	NaN	Clinopyroxene	0.955462	0.122916	Amphibole	0.043007	Augite
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
93026	48.531633	1.086635	6.924229	5.679027	0.897636	13.753107	20.414549	0.316232	0.073446	0.035749	NaN	NaN	Clinopyroxene	0.926149	0.175677	Garnet	0.032151	Diopside
93027	46.290230	1.380040	7.945115	7.235223	0.274338	11.797178	19.396081	0.955456	-0.142047	0.399850	NaN	NaN	Clinopyroxene	0.887151	0.191542	Glass	0.055837	Diopside
93028	48.117886	0.149082	8.696416	5.671712	0.169030	13.621192	20.402754	1.625494	0.058249	-0.174842	NaN	NaN	Clinopyroxene	0.971680	0.096162	Melilite	0.015869	Diopside
93029	46.544554	0.287509	7.870030	6.250091	0.117071	12.650849	19.429830	-0.016576	0.135432	-0.335588	NaN	NaN	Clinopyroxene	0.868953	0.225773	Garnet	0.056096	Diopside
93030	45.316298	0.119567	7.955299	6.887216	0.233471	11.919362	19.708412	0.945937	0.133037	0.141968	NaN	NaN	Clinopyroxene	0.951828	0.099333	Amphibole	0.025499	Diopside

14067 rows × 18 columns

'Oxides:'

	SiO2	TiO2	Al2O3	FeOt	MnO	MgO	CaO	Na2O	K2O	Cr2O3	P2O5	ZrO2	Predict_Mineral	Prediction_Score	Prediction_Score_Sigma	Second_Predict_Mineral	Second_Prediction_Score	Submineral
596	14.170745	0.458430	47.168928	17.851108	0.157325	8.807434	2.364795	1.613537	0.231962	0.909729	NaN	NaN	Oxide	0.974216	0.141011	Chlorite	2.158251e-02	Spinel_Group
597	-0.541988	-0.226758	59.577095	22.761509	0.261160	16.543655	0.178530	-0.327176	-0.002519	0.596073	NaN	NaN	Oxide	0.999999	0.000000	Chlorite	6.135478e-07	Spinel_Group
598	0.328607	0.192199	60.074062	22.531151	-0.581629	16.206175	-0.071793	0.404589	-0.041741	1.762033	NaN	NaN	Oxide	0.999997	0.000000	Rhombohedral_Oxides	3.225375e-06	Spinel_Group
599	0.419032	-0.416014	59.597162	20.757558	-0.198876	16.481142	-0.157663	0.094932	0.114364	1.412415	NaN	NaN	Oxide	0.999999	0.000244	Rhombohedral_Oxides	6.114131e-07	Spinel_Group
879	-0.283622	-0.329943	58.807763	21.683161	0.131220	15.814531	-0.270460	0.445707	0.113306	0.450210	NaN	NaN	Oxide	0.999537	0.003239	Chlorite	4.634696e-04	Spinel_Group
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
92992	-0.972727	0.199933	61.268541	19.730463	0.777305	16.862860	0.416785	0.532657	0.209819	1.072017	NaN	NaN	Oxide	0.999999	0.000244	Chlorite	5.719100e-07	Spinel_Group
92993	-0.754471	-0.316257	58.986714	21.457833	0.307136	14.984843	-0.049067	-0.382976	-0.506373	0.888524	NaN	NaN	Oxide	0.961885	0.186700	Chlorite	3.811476e-02	Spinel_Group
92994	-0.476960	0.326008	59.845682	20.725138	0.091979	16.190988	0.066551	0.394945	0.127746	0.985187	NaN	NaN	Oxide	0.999744	0.001777	Rhombohedral_Oxides	2.553134e-04	Spinel_Group
92995	5.724564	0.254216	55.985948	19.883002	-0.329296	15.163671	0.563598	1.013702	0.281037	1.044170	NaN	NaN	Oxide	0.996396	0.024116	Chlorite	3.439376e-03	Spinel_Group
92999	-0.771758	-0.311966	61.128154	23.235195	-0.302034	16.418883	0.084923	0.164328	0.299144	1.515729	NaN	NaN	Oxide	0.980000	0.140000	Nepheline	1.999995e-02	Spinel_Group

1902 rows × 18 columns

Plot these feldspars, pyroxenes, and spinels!

# Use FeldsparClassifier to examine at the component space (XAn, XAb, XOr)
fspar_comp = mm.FeldsparClassifier(fspars).calculate_components()
display(fspar_comp)

# Use FeldsparClassifier to plot up these data.
fig = mm.FeldsparClassifier(fspars).plot()

	Sample	SiO2	TiO2	Al2O3	FeOt	MnO	MgO	CaO	Na2O	K2O	...	Prediction_Score	Prediction_Score_Sigma	Second_Predict_Mineral	Second_Prediction_Score	Cation_Sum	M_site	T_site	An	Ab	Or
0	NaN	49.263735	0.246671	31.267656	0.000000	0.000000	0.000000	11.384471	3.906129	0.287173	...	0.998627	0.005443	Muscovite	0.001223	4.999206	0.944346	4.038491	0.605725	0.376083	0.018192
1	NaN	52.781465	1.223476	29.213528	1.751936	0.000000	0.353991	13.811718	3.613949	0.326185	...	0.999909	0.000345	Titanite	0.000041	5.005620	0.989984	3.885500	0.665953	0.315321	0.018726
2	NaN	49.582496	0.000000	31.217623	0.669667	0.312345	0.135481	13.362425	2.559899	0.404478	...	0.999078	0.002773	Muscovite	0.000342	4.973277	0.918395	4.007258	0.723219	0.250716	0.026065
3	NaN	49.766286	0.000000	30.748715	0.000000	0.052490	0.000000	13.924789	3.722859	0.031619	...	0.987317	0.083828	Leucite	0.007627	5.019679	1.028596	3.989021	0.672721	0.325460	0.001819
4	NaN	50.775667	0.283931	29.405951	0.363973	0.161280	0.434312	12.246722	2.495524	0.260420	...	0.997432	0.005905	Glass	0.001690	4.923061	0.851499	3.977005	0.717332	0.264506	0.018162
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
92980	NaN	50.751835	0.053460	30.892806	0.040494	0.000000	0.181496	13.219259	4.255093	0.562793	...	0.998067	0.008614	Glass	0.001712	5.053574	1.055605	3.977727	0.612308	0.356653	0.031038
92981	NaN	47.055855	0.435257	28.540332	0.088852	0.000000	0.027772	10.548025	3.540654	0.550766	...	0.980982	0.046478	Glass	0.012173	4.987109	0.940703	4.023037	0.598951	0.363812	0.037237
92982	NaN	49.572227	0.371584	32.248626	0.873017	0.000000	0.130899	12.215584	4.380849	0.471066	...	0.999816	0.000423	Muscovite	0.000078	5.065717	1.008939	3.983479	0.590014	0.382896	0.027090
92983	NaN	49.374082	0.000000	32.780453	0.389758	0.339789	0.275332	13.295722	4.352765	0.428836	...	0.997028	0.018393	Leucite	0.001778	5.089961	1.051290	3.982672	0.613189	0.363263	0.023548
92984	NaN	47.506399	0.326601	30.147186	0.261584	0.000000	0.000000	12.209264	4.127070	0.188657	...	0.990222	0.040914	Glass	0.007537	5.045055	1.026253	3.996449	0.613466	0.375247	0.011287

32058 rows × 58 columns

# Use PyroxeneClassifier to examine at the component space (En, Wo, Fs). If sodic pyroxenes are also within this input, this will plot them up in the sodic pyroxene ternary
pxs_comp = mm.PyroxeneClassifier(pxs).calculate_components()
display(pxs_comp)

# Use PyroxeneClassifier to plot up these data.
fig = mm.PyroxeneClassifier(pxs).plot()

	Sample	SiO2	TiO2	Al2O3	FeOt	MnO	MgO	CaO	Na2O	K2O	...	EnFs	DiHd_2003	Di	Fe3_Wang21	Fe2_Wang21	Di_h	Hd_h	Aeg_h	Jd_h	En_h
49	NaN	46.703956	2.306829	8.091728	7.608155	0.179336	10.411849	18.795262	2.432519	0.532825	...	0.086829	0.666407	0.469393	0.089679	0.154568	0.595819	0.009059	0.115772	0.018559	0.034919
610	NaN	48.659501	1.695958	7.334367	7.926977	0.000000	10.416438	17.438423	2.285797	1.183322	...	0.107573	0.632821	0.443486	0.013317	0.240387	0.594261	0.014598	0.071088	0.104224	0.066707
616	NaN	48.778790	1.189971	2.368374	9.717480	0.000000	19.703388	16.359033	0.808957	0.195681	...	0.417242	0.575877	0.451074	0.155597	0.150052	0.457241	0.000000	0.033339	0.000000	0.323735
626	NaN	43.817565	1.304287	4.129765	5.646358	0.000000	11.766227	22.258578	1.406237	0.008721	...	0.015722	0.891255	0.702211	0.187016	0.008697	0.726986	0.114188	0.062239	0.000000	0.000000
900	NaN	48.447279	1.397534	3.416259	9.859029	0.064421	16.668941	17.343585	1.357627	0.303317	...	0.317609	0.617466	0.462866	0.180335	0.131760	0.522099	0.000000	0.068710	0.000000	0.213817
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
93026	NaN	48.531633	1.086635	6.924229	5.679027	0.897636	13.753107	20.414549	0.316232	0.073446	...	0.135029	0.681486	0.537139	-0.004870	0.183839	0.683252	0.000000	0.000000	0.000000	0.148472
93027	NaN	46.290230	1.380040	7.945115	7.235223	0.274338	11.797178	19.396081	0.955456	0.000000	...	0.131225	0.653965	0.481823	0.026500	0.208089	0.634156	0.000000	0.019312	0.000000	0.125825
93028	NaN	48.117886	0.149082	8.696416	5.671712	0.169030	13.621192	20.402754	1.625494	0.058249	...	0.125932	0.684517	0.551743	0.128591	0.048722	0.597250	0.000000	0.116397	0.000000	0.083585
93029	NaN	46.544554	0.287509	7.870030	6.250091	0.117071	12.650849	19.429830	0.000000	0.135432	...	0.162218	0.623032	0.485826	-0.043900	0.249511	0.639871	0.000000	0.000000	0.000000	0.155748
93030	NaN	45.316298	0.119567	7.955299	6.887216	0.233471	11.919362	19.708412	0.945937	0.133037	...	0.124045	0.687473	0.514851	0.084473	0.144554	0.623202	0.000000	0.071634	0.000000	0.092467

14067 rows × 86 columns

/home/docs/checkouts/readthedocs.org/user_builds/mineralml/conda/stable/lib/python3.9/site-packages/ternary/plotting.py:148: UserWarning: No data for colormapping provided via 'c'. Parameters 'vmin', 'vmax' will be ignored
  ax.scatter(xs, ys, vmin=vmin, vmax=vmax, **kwargs)

# Use OxideClassifier to examine at the component space.
oxs_comp = mm.OxideClassifier(oxs).calculate_components()
display(oxs_comp)

# Use OxideClassifier to plot up these data.
fig = mm.OxideClassifier(oxs).plot()

	SiO2	TiO2	Al2O3	FeOt	MnO	MgO	CaO	Na2O	K2O	Cr2O3	...	Cation_Sum	A_site	A_site_expanded	B_site	A_B_site	Fe_Ti	Fe3_prop	XR2	XR3	XTi
596	14.170745	0.458430	47.168928	17.851108	0.157325	8.807434	2.364795	1.613537	0.231962	0.909729	...	3.0	0.801846	0.805654	1.619091	2.424745	0.436489	0.000000	0.332263	0.663672	0.004065
597	0.000000	0.000000	59.577095	22.761509	0.261160	16.543655	0.178530	0.000000	0.000000	0.596073	...	3.0	0.989221	0.995001	2.000000	2.995001	0.497456	0.307035	0.332221	0.667779	0.000000
598	0.328607	0.192199	60.074062	22.531151	0.000000	16.206175	0.000000	0.404589	0.000000	1.762033	...	3.0	0.971775	0.971775	1.999551	2.971326	0.489140	0.280194	0.327051	0.671696	0.001254
599	0.419032	0.000000	59.597162	20.757558	0.000000	16.481142	0.000000	0.094932	0.114364	1.412415	...	3.0	0.993664	0.993664	1.986631	2.980295	0.456697	0.239540	0.333411	0.666589	0.000000
879	0.000000	0.000000	58.807763	21.683161	0.131220	15.814531	0.000000	0.445707	0.113306	0.450210	...	3.0	0.943236	0.946200	2.026900	2.973100	0.483592	0.349607	0.318254	0.681746	0.000000
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
92992	0.000000	0.199933	61.268541	19.730463	0.777305	16.862860	0.416785	0.532657	0.209819	1.072017	...	3.0	0.909033	0.925878	2.029424	2.955302	0.426036	0.370311	0.313294	0.685404	0.001302
92993	0.000000	0.000000	58.986714	21.457833	0.307136	14.984843	0.000000	0.000000	0.000000	0.888524	...	3.0	0.992954	1.000000	2.000000	3.000000	0.486031	0.201826	0.333333	0.666667	0.000000
92994	0.000000	0.326008	59.845682	20.725138	0.091979	16.190988	0.066551	0.394945	0.127746	0.985187	...	3.0	0.953691	0.955739	2.017969	2.973708	0.462154	0.299782	0.321396	0.676435	0.002168
92995	5.724564	0.254216	55.985948	19.883002	0.000000	15.163671	0.563598	1.013702	0.281037	1.044170	...	3.0	1.017350	1.017350	1.756704	2.774054	0.439204	0.016451	0.366738	0.631462	0.001800
92999	0.000000	0.000000	61.128154	23.235195	0.000000	16.418883	0.084923	0.164328	0.299144	1.515729	...	3.0	0.962063	0.962063	2.017811	2.979874	0.494272	0.313183	0.322854	0.677146	0.000000

1902 rows × 70 columns

/home/docs/checkouts/readthedocs.org/user_builds/mineralml/conda/stable/lib/python3.9/site-packages/ternary/plotting.py:148: UserWarning: No data for colormapping provided via 'c'. Parameters 'vmin', 'vmax' will be ignored
  ax.scatter(xs, ys, vmin=vmin, vmax=vmax, **kwargs)

You might note that the structure of these three ...Classifier classes is identical. That is intentional! mm.FeldsparClassifier, mm.PyroxeneClassifier, and mm.OxideClassifier all have calculate_components and plot methods embedded.

5. Plot chemical variation maps

We know the mineralogy now. What if you now want to inspect the chemical variation within the individual crystals? Pull the component maps created for each sample and plot this up with mm.plot_component_composite.

This function currently does this calculation for feldspars, pyroxenes, olivines, and amphibole. This can easily be expanded with all the stoichiometric mineral functions. Here, I will just show this for these common igneous phases.

# Inspect what’s available:
print(sorted(output["component_maps"].keys()))

#  Plot map highlighting internal compositional variation
fig = mm.plot_component_composite(output, # specify output from above
                                  title="09g3", # optionally add a title to this plot
                                  phases=['Plagioclase', 'Clinopyroxene', 'Orthopyroxene', 'Oxide', 'Olivine', 'Glass'], # phases of interest
                                  smooth_sigma=0.25, # add a Gaussian blur to smooth compositional data, usually turned off.
                                  scalebar_um=50, # define size of scalebar desired, in microns
                                  pixel_size_um=2, # define size of each pixel of scalebar, in microns
                                  scalebar_loc='upper right', # specify location for scalebar
                                  scalebar_col='black', # specify color for scalebar
                                  )

['Clinopyroxene.En', 'Clinopyroxene.Fs', 'Clinopyroxene.Wo', 'Feldspar.Ab', 'Feldspar.An', 'Feldspar.Or', 'Orthopyroxene.En', 'Orthopyroxene.Fs', 'Orthopyroxene.Wo']

One could alternatively use all the functions within mineralML.mapping to do these same things, in a more stepwise manner. Look through the documentation if you would like to use individual bits of this code.