diff --git a/sharp/metrics/tests/test_metrics.py b/sharp/metrics/tests/test_base.py
similarity index 100%
rename from sharp/metrics/tests/test_metrics.py
rename to sharp/metrics/tests/test_base.py
diff --git a/sharp/metrics/tests/test_fidelity.py b/sharp/metrics/tests/test_fidelity.py
new file mode 100644
index 0000000..87265f5
--- /dev/null
+++ b/sharp/metrics/tests/test_fidelity.py
@@ -0,0 +1,58 @@
+import numpy as np
+from sharp.metrics import outcome_fidelity
+
+
+def test_outcome_fidelity_no_target_pairs_rank_true():
+    contributions = np.array([[0.1, 0.2], [0.3, 0.4]])
+    target = np.array([0.5, 0.6])
+    avg_target = 0.55
+    target_max = 1
+    result = outcome_fidelity(contributions, target, avg_target, target_max, rank=True)
+    expected = (
+        1
+        - np.mean(np.abs(target - (avg_target - contributions.sum(axis=1))))
+        / target_max
+    )
+    assert np.isclose(result, expected)
+
+
+def test_outcome_fidelity_no_target_pairs_rank_false():
+    contributions = np.array([[0.1, 0.2], [0.3, 0.4]])
+    target = np.array([0.5, 0.6])
+    avg_target = 0.55
+    target_max = 1
+    result = outcome_fidelity(contributions, target, avg_target, target_max, rank=False)
+    expected = np.mean(
+        1 - np.abs(target - (avg_target + contributions.sum(axis=1))) / target_max
+    )
+    assert np.isclose(result, expected)
+
+
+def test_outcome_fidelity_with_target_pairs_rank_true():
+    contributions = np.array([[0.1, 0.2], [0.3, 0.4]])
+    target = np.array([0.5, 0.6])
+    avg_target = 0.55
+    target_max = 1
+    target_pairs = np.array([0.4, 0.7])
+    result = outcome_fidelity(
+        contributions, target, avg_target, target_max, target_pairs, rank=True
+    )
+    better_than = target < target_pairs
+    est_better_than = contributions.sum(axis=1) > 0
+    expected = (better_than == est_better_than).mean()
+    assert np.isclose(result, expected)
+
+
+def test_outcome_fidelity_with_target_pairs_rank_false():
+    contributions = np.array([[0.1, 0.2], [0.3, 0.4]])
+    target = np.array([0.5, 0.6])
+    avg_target = 0.55
+    target_max = 1
+    target_pairs = np.array([0.4, 0.7])
+    result = outcome_fidelity(
+        contributions, target, avg_target, target_max, target_pairs, rank=False
+    )
+    better_than = target > target_pairs
+    est_better_than = contributions.sum(axis=1) > 0
+    expected = (better_than == est_better_than).mean()
+    assert np.isclose(result, expected)
diff --git a/sharp/utils/tests/test_parallelize.py b/sharp/utils/tests/test_parallelize.py
new file mode 100644
index 0000000..baf20a3
--- /dev/null
+++ b/sharp/utils/tests/test_parallelize.py
@@ -0,0 +1,16 @@
+from sharp.utils._parallelize import parallel_loop
+
+
+def test_parallel_loop():
+    def square(x):
+        return x * x
+
+    iterable = range(10)
+
+    results = parallel_loop(square, iterable, n_jobs=2, progress_bar=False)
+    assert results == [x * x for x in iterable]
+
+    results = parallel_loop(
+        square, iterable, n_jobs=2, progress_bar=True, description="Test"
+    )
+    assert results == [x * x for x in iterable]
diff --git a/sharp/utils/tests/test_utils.py b/sharp/utils/tests/test_utils.py
new file mode 100644
index 0000000..d7c8a7d
--- /dev/null
+++ b/sharp/utils/tests/test_utils.py
@@ -0,0 +1,17 @@
+import pytest
+from sharp.utils._utils import _optional_import
+
+
+def test_optional_import_existing_module():
+    module = _optional_import("math")
+    assert module is not None
+
+
+def test_optional_import_non_existing_module():
+    with pytest.raises(ImportError):
+        _optional_import("non_existing_module")
+
+
+def test_optional_import_partial_module():
+    module = _optional_import("os.path")
+    assert module is not None
diff --git a/sharp/visualization/tests/test_aggregate.py b/sharp/visualization/tests/test_aggregate.py
new file mode 100644
index 0000000..0d671ac
--- /dev/null
+++ b/sharp/visualization/tests/test_aggregate.py
@@ -0,0 +1,45 @@
+import pytest
+import numpy as np
+import pandas as pd
+from sharp.visualization._aggregate import group_boxplot
+
+
+@pytest.fixture
+def sample_data():
+    X = pd.DataFrame(
+        {
+            "feature1": np.random.rand(100),
+            "feature2": np.random.rand(100),
+            "group": np.random.choice(["A", "B", "C", "D", "E"], 100),
+        }
+    )
+    y = np.random.rand(100)
+    contributions = np.random.rand(100, 2)
+    feature_names = ["feature1", "feature2"]
+    return X, y, contributions, feature_names
+
+
+def test_group_boxplot_group_by_bins(sample_data):
+    X, y, contributions, feature_names = sample_data
+    ax = group_boxplot(
+        X, y, contributions, feature_names=feature_names, group=5, show=False
+    )
+    assert ax is not None
+    assert len(ax.get_xticklabels()) == 5
+
+
+def test_group_boxplot_group_by_variable(sample_data):
+    X, y, contributions, feature_names = sample_data
+    ax = group_boxplot(
+        X, y, contributions, feature_names=feature_names, group="group", show=False
+    )
+    assert ax is not None
+    assert len(ax.get_xticklabels()) == len(X["group"].unique())
+
+
+def test_group_boxplot_show(sample_data):
+    X, y, contributions, feature_names = sample_data
+    ax = group_boxplot(
+        X, y, contributions, feature_names=feature_names, group=5, show=True
+    )
+    assert ax is None
diff --git a/sharp/visualization/tests/test_visualization.py b/sharp/visualization/tests/test_visualization.py
new file mode 100644
index 0000000..108f892
--- /dev/null
+++ b/sharp/visualization/tests/test_visualization.py
@@ -0,0 +1,87 @@
+import pytest
+import pandas as pd
+import numpy as np
+from sharp.utils._utils import _optional_import
+from sharp.visualization._visualization import ShaRPViz
+
+
+@pytest.fixture
+def mock_sharp():
+    """
+    Fixture to create a mock ShaRP object with dummy feature names.
+    """
+
+    class MockSharp:
+        def __init__(self):
+            self.feature_names_ = np.array(["Feature1", "Feature2", "Feature3"])
+            self.qoi="rank"
+            self.measure="shapley"
+
+    return MockSharp()
+
+
+def test_bar(mock_sharp):
+    """
+    Test the bar method of ShaRPViz.
+    """
+    sharpviz = ShaRPViz(mock_sharp)
+    scores = [0.1, 0.5, 0.4]
+
+    plt = _optional_import("matplotlib.pyplot")
+    fig, ax = plt.subplots(1, 1)
+
+    result_ax = sharpviz.bar(scores, ax=ax)
+    assert result_ax is not None
+    assert result_ax.get_ylabel() == "Contribution to QoI"
+    assert result_ax.get_xlabel() == "Features"
+    assert len(result_ax.patches) == len(scores)  # Number of bars matches scores
+
+
+def test_waterfall(mock_sharp):
+    """
+    Test the waterfall method of ShaRPViz.
+    """
+    sharpviz = ShaRPViz(mock_sharp)
+    contributions = [0.2, -0.1, 0.3]
+    feature_values = ["A", "B", "C"]
+    mean_target_value = 1.0
+
+    result = sharpviz.waterfall(
+        contributions=contributions,
+        feature_values=feature_values,
+        mean_target_value=mean_target_value,
+    )
+    assert result is not None
+
+
+def test_box(mock_sharp):
+    """
+    Test the box method of ShaRPViz.
+    """
+    sharpviz = ShaRPViz(mock_sharp)
+    X = pd.DataFrame(
+        np.random.randn(100, 3), columns=["Feature1", "Feature2", "Feature3"]
+    )
+    y = np.random.choice([0, 1], size=100)
+    contributions = pd.DataFrame(
+        np.random.randn(100, 3), columns=["Feature1", "Feature2", "Feature3"]
+    )
+
+    plt = _optional_import("matplotlib.pyplot")
+    fig, ax = plt.subplots(1, 1)
+
+    result = sharpviz.box(X, y, contributions, ax=ax)
+    assert result is not None
+
+
+def test_bar_without_ax(mock_sharp):
+    """
+    Test the bar method without providing an ax.
+    """
+    sharpviz = ShaRPViz(mock_sharp)
+    scores = [0.3, 0.5, 0.2]
+
+    result_ax = sharpviz.bar(scores)
+    assert result_ax is not None
+    assert result_ax.get_ylabel() == "Contribution to QoI"
+    assert result_ax.get_xlabel() == "Features"
diff --git a/sharp/visualization/tests/test_waterfall.py b/sharp/visualization/tests/test_waterfall.py
new file mode 100644
index 0000000..1920f65
--- /dev/null
+++ b/sharp/visualization/tests/test_waterfall.py
@@ -0,0 +1,57 @@
+import pytest
+import pandas as pd
+import numpy as np
+from sharp.visualization._waterfall import format_value, _waterfall
+
+
+@pytest.mark.parametrize(
+    "value, format_str, expected",
+    [
+        (123.45000, "%.2f", "123.45"),
+        (123.00000, "%.2f", "123"),
+        (-123.45000, "%.2f", "\u2212123.45"),
+        (-123.00000, "%.2f", "\u2212123"),
+        (0.00000, "%.2f", "0"),
+        ("123.45", "%.2f", "123.45"),
+        ("-123.45", "%.2f", "\u2212123.45"),
+    ],
+)
+def test_format_value(value, format_str, expected):
+    assert format_value(value, format_str) == expected
+
+
+@pytest.fixture
+def shap_values():
+    return {
+        "base_values": 0.5,
+        "features": np.array([1.0, 2.0, 3.0]),
+        "feature_names": ["feature1", "feature2", "feature3"],
+        "values": pd.Series([0.1, -0.2, 0.3]),
+    }
+
+
+def test_waterfall_plot(shap_values):
+    fig = _waterfall(shap_values, max_display=3, show=False)
+    assert fig is not None
+    assert len(fig.axes) > 0
+
+
+@pytest.mark.parametrize("max_display", [1, 2, 3])
+def test_waterfall_max_display(shap_values, max_display):
+    fig = _waterfall(shap_values, max_display=max_display, show=False)
+    assert fig is not None
+    assert (
+        len(fig.axes[0].patches) == max_display * 2
+    )  # Each feature has two patches (positive and negative)
+
+
+def test_waterfall_no_features():
+    shap_values = {
+        "base_values": 0.5,
+        "features": None,
+        "feature_names": ["feature1", "feature2", "feature3"],
+        "values": pd.Series([0.1, -0.2, 0.3]),
+    }
+    fig = _waterfall(shap_values, max_display=3, show=False)
+    assert fig is not None
+    assert len(fig.axes) > 0