Load MAREoS dataset

Imports

from pathlib import Path

import pandas as pd

from uniharmony.datasets import load_MAREoS

We can call the helper funtion to load all the dataset (aprox 3MB). The files will be stored in the cache, so we don’t have to worry about them

datasets = load_MAREoS()

2026-05-18 13:04:33 [info     ] MAREoS datasets already exist at: /home/runner/.cache/uniharmony/MAREoS
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_simple1_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_simple2_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_interaction1_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_interaction2_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/true_simple1_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/true_simple2_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/true_interaction1_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/true_interaction2_data.csv

Exploration

Let’s explore now how the datasets looks like

print(datasets.keys())

dict_keys(['eos_simple1', 'eos_simple2', 'eos_interaction1', 'eos_interaction2', 'true_simple1', 'true_simple2', 'true_interaction1', 'true_interaction2'])

We have now all the datasets in a dictionary. There is a total of 8 datasets.

# Select one dataset and explore what is inside the dictionary
dataset = datasets["eos_simple1"]
print(dataset.keys())

dict_keys(['X', 'y', 'sites', 'covs', 'folds'])

Let’s unpack what is inside the keys. This is the typical way you can use the dataset for further downstream analysis.

X = dataset["X"]
y = dataset["y"]

print(f"Load X with shape:{X.shape} and y:{y.shape}")

Load X with shape:(1001, 14) and y:(1001,)

Variations

# You can use the helper function to only return a part of the datasets
datasets = load_MAREoS(effects="eos")
print(datasets.keys())

2026-05-18 13:04:33 [info     ] MAREoS datasets already exist at: /home/runner/.cache/uniharmony/MAREoS
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_simple1_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_simple2_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_interaction1_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_interaction2_data.csv
dict_keys(['eos_simple1', 'eos_simple2', 'eos_interaction1', 'eos_interaction2'])

datasets = load_MAREoS(effects="eos", effect_types="simple")
print(datasets.keys())

2026-05-18 13:04:33 [info     ] MAREoS datasets already exist at: /home/runner/.cache/uniharmony/MAREoS
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_simple1_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_simple2_data.csv
dict_keys(['eos_simple1', 'eos_simple2'])

datasets = load_MAREoS(effects="eos", effect_types="simple", effect_examples="1")
print(datasets.keys())

2026-05-18 13:04:33 [info     ] MAREoS datasets already exist at: /home/runner/.cache/uniharmony/MAREoS
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_simple1_data.csv
dict_keys(['eos_simple1'])

Returning the dataset as DataFrame allows to see the simulated areas You can chose to load the dataset as pandas.DataFrame, with has the simulated areas of the brain.

datasets = load_MAREoS(effects="eos", effect_types="simple", effect_examples="1", as_numpy=False)
dataset = datasets["eos_simple1"]["X"]
# Show only 5 columns
pd.set_option("display.max_columns", 8)
dataset.head()

2026-05-18 13:04:33 [info     ] MAREoS datasets already exist at: /home/runner/.cache/uniharmony/MAREoS
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/.cache/uniharmony/MAREoS/public_datasets/eos_simple1_data.csv

	Lthal	Rthal	Lcaud	Rcaud	...	Lamyg	Ramyg	Laccumb	Raccumb
id
1	8895.369099	8383.870372	3803.558492	4357.165963	...	1563.622497	1854.229137	698.637972	701.906213
2	8679.346875	6654.136742	3924.041654	3745.063498	...	1511.933005	1020.152948	709.090077	534.448106
3	9191.801201	7159.776871	3444.265568	3158.455008	...	1761.132128	1114.841164	785.199200	717.806882
4	7531.473405	6694.021219	4984.063517	4689.035649	...	1774.472741	742.652391	1104.007368	769.837240
5	7070.478721	5575.244389	3285.175734	2234.129050	...	1722.570593	1414.669078	1000.597680	440.375965

5 rows × 14 columns

Load the dataset in a user determine folder We could also want to see the csv files in a folder, we could pass a directory for the function to save the data Let’s pass a directory inside the repository. We will use a relative path from this example to look for appropiated path

data_dir = Path().resolve().parents[1] / "src" / "uniharmony" / "datasets" / "data"
datasets = load_MAREoS(data_dir=data_dir)

Unzipping contents of '/home/runner/.cache/uniharmony/public_datasets.zip' to '/home/runner/work/UniHarmony/UniHarmony/src/uniharmony/datasets/data/MAREoS'
2026-05-18 13:04:33 [info     ] MAREoS datasets downloaded: 16 CSV files in /home/runner/work/UniHarmony/UniHarmony/src/uniharmony/datasets/data/MAREoS
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/work/UniHarmony/UniHarmony/src/uniharmony/datasets/data/MAREoS/public_datasets/eos_simple1_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/work/UniHarmony/UniHarmony/src/uniharmony/datasets/data/MAREoS/public_datasets/eos_simple2_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/work/UniHarmony/UniHarmony/src/uniharmony/datasets/data/MAREoS/public_datasets/eos_interaction1_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/work/UniHarmony/UniHarmony/src/uniharmony/datasets/data/MAREoS/public_datasets/eos_interaction2_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/work/UniHarmony/UniHarmony/src/uniharmony/datasets/data/MAREoS/public_datasets/true_simple1_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/work/UniHarmony/UniHarmony/src/uniharmony/datasets/data/MAREoS/public_datasets/true_simple2_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/work/UniHarmony/UniHarmony/src/uniharmony/datasets/data/MAREoS/public_datasets/true_interaction1_data.csv
2026-05-18 13:04:33 [info     ] Getting data file: /home/runner/work/UniHarmony/UniHarmony/src/uniharmony/datasets/data/MAREoS/public_datasets/true_interaction2_data.csv