Stratified splitting of pandas dataframe in training, validation and test set

np.array_split

If you want to generalise to n splits, np.array_split is your friend (it works with DataFrames well).

fractions = np.array([0.6, 0.2, 0.2])
# shuffle your input
df = df.sample(frac=1) 
# split into 3 parts
train, val, test = np.array_split(
    df, (fractions[:-1].cumsum() * len(df)).astype(int))

---

train_test_split

A windy solution using train_test_split for stratified splitting.

y = df.pop('diagnosis').to_frame()
X = df

X_train, X_test, y_train, y_test = train_test_split(
        X, y,stratify=y, test_size=0.4)

X_test, X_val, y_test, y_val = train_test_split(
        X_test, y_test, stratify=y_test, test_size=0.5)

Where X is a DataFrame of your features, and y is a single-columned DataFrame of your labels.