diff --git a/mixnet-rs/analysis/.gitignore b/mixnet-rs/analysis/.gitignore
new file mode 100644
index 0000000..21d0b89
--- /dev/null
+++ b/mixnet-rs/analysis/.gitignore
@@ -0,0 +1 @@
+.venv/
diff --git a/mixnet-rs/analysis/coeff.py b/mixnet-rs/analysis/coeff.py
new file mode 100644
index 0000000..39c2461
--- /dev/null
+++ b/mixnet-rs/analysis/coeff.py
@@ -0,0 +1,96 @@
+import argparse
+
+import matplotlib.pyplot as plt
+import pandas as pd
+
+from common import MARKERS, X_FIELDS
+
+
+def analyze(path: str):
+    data = pd.read_csv(path)
+    for x_field in X_FIELDS:
+        analyze_versus(data, x_field)
+
+
+def analyze_versus(data: pd.DataFrame, x_field: str):
+    # Group by both x_field and queue_type, then select the row with the largest paramset for each group
+    max_paramset_data = data.loc[
+        data.groupby([x_field, "queue_type"])["paramset"].idxmax()
+    ]
+
+    all_fields = [
+        ["strong_coeff_min", "casual_coeff_min", "weak_coeff_min"],
+        ["strong_coeff_mean", "casual_coeff_mean", "weak_coeff_mean"],
+        ["strong_coeff_max", "casual_coeff_max", "weak_coeff_max"],
+    ]
+
+    # Display the plots
+    fig, ax = plt.subplots(3, 3, figsize=(20, 10))
+    for ax_col, fields in enumerate(all_fields):
+        max_y = 0
+        for field in fields:
+            max_y = max(max_y, max_paramset_data[field].max())
+        max_y = min(max_y, 5000)  # hard limit
+
+        for ax_row, field in enumerate(fields):
+            for queue_type in max_paramset_data["queue_type"].unique():
+                queue_data = max_paramset_data[
+                    max_paramset_data["queue_type"] == queue_type
+                ]
+                x_values = queue_data[x_field]
+                y_values = queue_data[field]
+                ax[ax_row][ax_col].plot(
+                    x_values,
+                    y_values,
+                    label=queue_type,
+                    marker=MARKERS[queue_type],
+                )
+
+            ax[ax_row][ax_col].set_title(f"{field} vs {x_field}", fontsize=10)
+            ax[ax_row][ax_col].set_xlabel(x_field)
+            ax[ax_row][ax_col].set_ylabel(field)
+            if ax_row == 0 and ax_col == len(all_fields) - 1:
+                ax[ax_row][ax_col].legend(bbox_to_anchor=(1, 1), loc="upper left")
+            ax[ax_row][ax_col].grid(True)
+            ax[ax_row][ax_col].set_ylim(-10, max_y * 1.05)
+
+    plt.tight_layout()
+
+    # Display the table of values
+    # Create a table with x_field, queue_type, and coefficients
+    flatten_fields = [
+        field for zipped_fields in zip(*all_fields) for field in zipped_fields
+    ]
+    columns = [x_field, "queue_type"] + flatten_fields
+    table_data = max_paramset_data[columns].sort_values(by=[x_field, "queue_type"])
+    # Prepare to display values with only 2 decimal places
+    table_data[flatten_fields] = table_data[flatten_fields].applymap(
+        lambda x: f"{x:.2f}"
+    )
+    # Display the table as a separate subplot
+    fig_table, ax_table = plt.subplots(
+        figsize=(len(columns) * 1.8, len(table_data) * 0.3)
+    )
+    ax_table.axis("off")  # Turn off the axis
+    table = ax_table.table(
+        cellText=table_data.values,
+        colLabels=table_data.columns,
+        cellLoc="center",
+        loc="center",
+    )
+    table.auto_set_font_size(False)
+    table.set_fontsize(10)
+    table.scale(1, 1.5)
+    for i in range(len(table_data.columns)):
+        table.auto_set_column_width(i)
+
+    plt.show()
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(
+        description="Aggregate the results of all paramsets of an experiment"
+    )
+    parser.add_argument("path", type=str, help="dir path")
+    args = parser.parse_args()
+    analyze(args.path)
diff --git a/mixnet-rs/analysis/common.py b/mixnet-rs/analysis/common.py
new file mode 100644
index 0000000..8261416
--- /dev/null
+++ b/mixnet-rs/analysis/common.py
@@ -0,0 +1,15 @@
+MARKERS = {
+    "NonMix": "o",
+    "PureCoinFlipping": "s",
+    "PureRandomSampling": "^",
+    "PermutedCoinFlipping": "v",
+    "NoisyCoinFlipping": "*",
+    "NoisyCoinFlippingRandomRelease": "d",
+}
+
+X_FIELDS = [
+    "num_mixes",
+    "peering_degree",
+    "sender_data_msg_prob",
+    "mix_data_msg_prob",
+]
diff --git a/mixnet-rs/analysis/latency.py b/mixnet-rs/analysis/latency.py
new file mode 100644
index 0000000..abfd5e4
--- /dev/null
+++ b/mixnet-rs/analysis/latency.py
@@ -0,0 +1,81 @@
+import argparse
+
+import matplotlib.pyplot as plt
+import pandas as pd
+
+from common import MARKERS, X_FIELDS
+
+
+def analyze(path: str):
+    data = pd.read_csv(path)
+    for x_field in X_FIELDS:
+        analyze_versus(data, x_field)
+
+
+def analyze_versus(data: pd.DataFrame, x_field: str):
+    # Group by both x_field and queue_type, then select the row with the largest paramset for each group
+    max_paramset_data = data.loc[
+        data.groupby([x_field, "queue_type"])["paramset"].idxmax()
+    ]
+
+    fields = ["latency_min", "latency_mean", "latency_max"]
+
+    # Display the plots
+    fig, ax = plt.subplots(1, 3, figsize=(20, 4))
+    for ax_col, field in enumerate(fields):
+        for queue_type in max_paramset_data["queue_type"].unique():
+            queue_data = max_paramset_data[
+                max_paramset_data["queue_type"] == queue_type
+            ]
+            x_values = queue_data[x_field]
+            y_values = queue_data[field]
+            ax[ax_col].plot(
+                x_values,
+                y_values,
+                label=queue_type,
+                marker=MARKERS[queue_type],
+            )
+
+        ax[ax_col].set_title(f"{field} vs {x_field}", fontsize=10)
+        ax[ax_col].set_xlabel(x_field)
+        ax[ax_col].set_ylabel(field)
+        if ax_col == len(fields) - 1:
+            ax[ax_col].legend(bbox_to_anchor=(1, 1), loc="upper left")
+        ax[ax_col].grid(True)
+        ax[ax_col].set_ylim(bottom=0)
+
+    plt.tight_layout()
+
+    # Display the table of values
+    # Create a table with x_field, queue_type, and coefficients
+    columns = [x_field, "queue_type"] + fields
+    table_data = max_paramset_data[columns].sort_values(by=[x_field, "queue_type"])
+    # Prepare to display values with only 2 decimal places
+    table_data[fields] = table_data[fields].applymap(lambda x: f"{x:.2f}")
+    # Display the table as a separate subplot
+    fig_table, ax_table = plt.subplots(
+        figsize=(len(columns) * 1.8, len(table_data) * 0.3)
+    )
+    ax_table.axis("off")  # Turn off the axis
+    table = ax_table.table(
+        cellText=table_data.values,
+        colLabels=table_data.columns,
+        cellLoc="center",
+        loc="center",
+    )
+    table.auto_set_font_size(False)
+    table.set_fontsize(10)
+    table.scale(1, 1.5)
+    for i in range(len(table_data.columns)):
+        table.auto_set_column_width(i)
+
+    plt.show()
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(
+        description="Aggregate the results of all paramsets of an experiment"
+    )
+    parser.add_argument("path", type=str, help="dir path")
+    args = parser.parse_args()
+    analyze(args.path)
diff --git a/mixnet-rs/analysis/pyrightconfig.json b/mixnet-rs/analysis/pyrightconfig.json
new file mode 100644
index 0000000..9cd2c6d
--- /dev/null
+++ b/mixnet-rs/analysis/pyrightconfig.json
@@ -0,0 +1,7 @@
+{
+  "venvPath": ".",
+  "venv": ".venv",
+  "typeCheckingMode": "basic",
+  "reportMissingTypeStubs": "none",
+  "ignore": ["reportMissingTypeStubs"]
+}
diff --git a/mixnet-rs/analysis/requirements.txt b/mixnet-rs/analysis/requirements.txt
new file mode 100644
index 0000000..94f4c83
--- /dev/null
+++ b/mixnet-rs/analysis/requirements.txt
@@ -0,0 +1,2 @@
+matplotlib==3.9.2
+pandas==2.2.2