Update app.py

app.py

@@ -2,121 +2,145 @@ import streamlit as st
 import pandas as pd
 import numpy as np
 
-# Function to generate synthetic data
-def generate_synthetic_data(num_members):
-    unique_ids = [f"MEM_{i:05d}" for i in range(1, num_members + 1)]
-    primary_keys = [f"PPK_{i:05d}" for i in range(1, num_members + 1)]
-
-    # Synthetic Enrollments
-    enrollments_data = {
-        "MEM_AGE": np.random.randint(18, 80, num_members),
-        "MEM_MSA_NAME": np.random.choice(["DETROIT", "HONOLULU", "LOS ANGELES"], num_members),
-        "MEM_STAT": np.random.choice(["ACTIVE", "INACTIVE"], num_members),
-        "MEMBER_ID": unique_ids,
-        "PRIMARY_PERSON_KEY": primary_keys,
-        "PAYER_LOB": np.random.choice(["MEDICAID", "COMMERCIAL", "MEDICARE"], num_members),
-        "PAYER_TYPE": np.random.choice(["PPO", "HMO"], num_members),
-        "PRIMARY_CHRONIC_CONDITION_ROLLUP_DESC": np.random.choice(["Cancer", "Diabetes", "Hypertension"], num_members),
-        "Count of PRIMARY_CHRONIC_CONDITION_ROLLUP_ID": np.random.randint(1, 5, num_members),
-        "PROD_TYPE": np.random.choice(["DENTAL", "VISION", "MEDICAL"], num_members),
-        "RELATION": np.random.choice(["SUBSCRIBER", "DEPENDENT"], num_members),
-        "Sum of QTY_MM_DN": np.random.randint(0, 10, num_members),
-        "Sum of QTY_MM_MD": np.random.randint(0, 10, num_members),
-        "Sum of QTY_MM_RX": np.random.randint(0, 10, num_members),
-        "Sum of QTY_MM_VS": np.random.randint(0, 10, num_members),
-        "YEARMO": np.random.randint(202201, 202412, num_members),
-    }
-    enrollments_df = pd.DataFrame(enrollments_data)
-
-    # Synthetic Members
-    members_data = {
-        "MEM_ETHNICITY": np.random.choice(["Hispanic", "Non-Hispanic", None], num_members),
-        "MEM_GENDER": ["F"] * num_members,  # Ensuring all members are female
-        "MEM_MSA_NAME": enrollments_data["MEM_MSA_NAME"],
-        "MEM_RACE": np.random.choice(["White", "Black", "Asian", None], num_members),
-        "MEM_STATE": np.random.choice(["MI", "HI", "CA"], num_members),
-        "MEM_ZIP3": np.random.randint(100, 999, num_members),
-        "MEMBER_ID": unique_ids,
-        "PRIMARY_PERSON_KEY": primary_keys,
-    }
-    members_df = pd.DataFrame(members_data)
-
-    # Synthetic Providers
-    providers_data = {
-        "PROV_CLINIC_STATE": np.random.choice(["MI", "HI", "CA"], num_members),
-        "PROV_CLINIC_ZIP": np.random.randint(10000, 99999, num_members),
-        "PROV_KEY": [f"PK_{i:05d}" for i in range(1, num_members + 1)],
-        "PROV_NPI_ORG": np.random.randint(1, 50, num_members),
-        "PROV_TAXONOMY": np.random.choice(["208100000X", "207RE0101X"], num_members),
-        "PROV_TYPE": np.random.choice(["Type1", "Type2"], num_members),
-    }
-    providers_df = pd.DataFrame(providers_data)
-
-    # Synthetic Services
-    services_data = {
-        "MEMBER_ID": unique_ids,
-        "PRIMARY_PERSON_KEY": primary_keys,
-        "Sum of AMT_ALLOWED": np.random.uniform(1000, 10000, num_members),
-        "Sum of AMT_BILLED": np.random.uniform(1000, 15000, num_members),
-        "Count of AMT_PAID": np.random.randint(1, 5, num_members),
-        "ATT_PROV_KEY": np.random.choice(providers_data["PROV_KEY"], num_members),
-        "BILL_PROV_KEY": np.random.choice(providers_data["PROV_KEY"], num_members),
-        "CLAIM_IN_NETWORK": np.random.choice(["Y", "N", None], num_members),
-        "RELATION": enrollments_data["RELATION"],
-        "SERVICE_SETTING": np.random.choice(["OUTPATIENT", "INPATIENT"], num_members),
-        "Sum of SERVICE_LINE": np.random.randint(1, 10, num_members),
-        "Sum of SV_UNITS": np.random.randint(1, 100, num_members),
-        "YEARMO": enrollments_data["YEARMO"],
-    }
-    services_df = pd.DataFrame(services_data)
-
-    return enrollments_df, members_df, providers_df, services_df
-
-
-# Streamlit App
-st.title("Synthetic Medical Billing Data Generator")
-
-# Slider for number of members
-num_members = st.slider("Select number of unique members:", min_value=10, max_value=1000, step=10, value=100)
-
-# Generate synthetic data
-enrollments_df, members_df, providers_df, services_df = generate_synthetic_data(num_members)
-
-# Display dataframes
-st.subheader("Preview of Generated Data")
-st.write("Enrollments Data")
-st.dataframe(enrollments_df.head())
-st.write("Members Data")
-st.dataframe(members_df.head())
-st.write("Providers Data")
-st.dataframe(providers_df.head())
-st.write("Services Data")
-st.dataframe(services_df.head())
-
-# Allow downloading the generated files
-st.subheader("Download Synthetic Data")
-st.download_button(
-    label="Download Enrollments Data",
-    data=enrollments_df.to_csv(index=False),
-    file_name="Synthetic_Enrollments.csv",
-    mime="text/csv",
-)
-st.download_button(
-    label="Download Members Data",
-    data=members_df.to_csv(index=False),
-    file_name="Synthetic_Members.csv",
-    mime="text/csv",
-)
-st.download_button(
-    label="Download Providers Data",
-    data=providers_df.to_csv(index=False),
-    file_name="Synthetic_Providers.csv",
-    mime="text/csv",
-)
-st.download_button(
-    label="Download Services Data",
-    data=services_df.to_csv(index=False),
-    file_name="Synthetic_Services.csv",
-    mime="text/csv",
-)
+# Seed for reproducibility
+np.random.seed(42)
 
+# Function to generate synthetic breast cancer data
+def generate_breast_cancer_data(num_patients=100):
+    patient_ids = [f"BC_{i+1:05d}" for i in range(num_patients)]
+    ages = []
+    menopausal_status = []
+    tumor_sizes = []
+    lymph_nodes = []
+    grades = []
+    stages = []
+    er_status = []
+    pr_status = []
+    her2_status = []
+    ki67_level = []
+    tnbc_status = []
+    brca_mutation = []
+    overall_health = []
+    genomic_score = []
+    treatment = []
+
+    for i in range(num_patients):
+        age = int(np.random.normal(60, 10))
+        age = max(30, min(age, 80))
+        ages.append(age)
+
+        menopausal = "Post-menopausal" if age >= 50 else "Pre-menopausal"
+        menopausal_status.append(menopausal)
+
+        tumor_size = round(np.random.lognormal(mean=0.7, sigma=0.5), 2)
+        tumor_sizes.append(tumor_size)
+
+        lymph_node = (
+            "Positive"
+            if (tumor_size > 2.0 and np.random.rand() < 0.6)
+            or (tumor_size <= 2.0 and np.random.rand() < 0.3)
+            else "Negative"
+        )
+        lymph_nodes.append(lymph_node)
+
+        grade = np.random.choice([1, 2, 3], p=[0.1, 0.4, 0.5] if tumor_size > 2.0 else [0.3, 0.5, 0.2])
+        grades.append(grade)
+
+        if tumor_size <= 2.0 and lymph_node == "Negative":
+            stage = "I"
+        elif (tumor_size > 2.0 and tumor_size <= 5.0) and lymph_node == "Negative":
+            stage = "II"
+        elif lymph_node == "Positive" or tumor_size > 5.0:
+            stage = "III"
+        else:
+            stage = "II"
+        if np.random.rand() < 0.05:
+            stage = "IV"
+        stages.append(stage)
+
+        er = np.random.choice(["Positive", "Negative"], p=[0.75, 0.25])
+        pr = "Positive" if er == "Positive" and np.random.rand() > 0.1 else "Negative"
+        er_status.append(er)
+        pr_status.append(pr)
+
+        her2 = np.random.choice(["Positive", "Negative"], p=[0.3, 0.7] if grade == 3 else [0.15, 0.85])
+        her2_status.append(her2)
+
+        ki67 = "High" if grade == 3 and np.random.rand() < 0.8 else "Low"
+        ki67_level.append(ki67)
+
+        tnbc = "Positive" if er == "Negative" and pr == "Negative" and her2 == "Negative" else "Negative"
+        tnbc_status.append(tnbc)
+
+        brca = "Positive" if (tnbc == "Positive" or age < 40) and np.random.rand() < 0.2 else "Negative"
+        brca_mutation.append(brca)
+
+        health = "Good" if age < 65 and np.random.rand() < 0.9 else "Poor"
+        overall_health.append(health)
+
+        recurrence_score = (
+            np.random.choice(["Low", "Intermediate", "High"], p=[0.6, 0.3, 0.1])
+            if er == "Positive" and her2 == "Negative"
+            else "N/A"
+        )
+        genomic_score.append(recurrence_score)
+
+        if stage in ["I", "II"]:
+            if tnbc == "Positive":
+                treat = "Surgery, Chemotherapy, and Radiation Therapy"
+            elif er == "Positive" and recurrence_score != "N/A":
+                if recurrence_score == "High":
+                    treat = "Surgery, Chemotherapy, Hormone Therapy, and Radiation Therapy"
+                elif recurrence_score == "Intermediate":
+                    treat = "Surgery, Consider Chemotherapy, Hormone Therapy, and Radiation Therapy"
+                else:
+                    treat = "Surgery, Hormone Therapy, and Radiation Therapy"
+            elif her2 == "Positive":
+                treat = "Surgery, HER2-Targeted Therapy, Chemotherapy, and Radiation Therapy"
+            else:
+                treat = "Surgery, Chemotherapy, and Radiation Therapy"
+        elif stage == "III":
+            treat = (
+                "Neoadjuvant Chemotherapy, Surgery, Radiation Therapy"
+                + (", HER2-Targeted Therapy" if her2 == "Positive" else "")
+                + (", Hormone Therapy" if er == "Positive" else "")
+            )
+        else:
+            treat = "Systemic Therapy (Palliative Care)"
+        treatment.append(treat)
+
+    return pd.DataFrame(
+        {
+            "Patient ID": patient_ids,
+            "Age": ages,
+            "Menopausal Status": menopausal_status,
+            "Tumor Size (cm)": tumor_sizes,
+            "Lymph Node Involvement": lymph_nodes,
+            "Tumor Grade": grades,
+            "Tumor Stage": stages,
+            "ER Status": er_status,
+            "PR Status": pr_status,
+            "HER2 Status": her2_status,
+            "Ki-67 Level": ki67_level,
+            "TNBC Status": tnbc_status,
+            "BRCA Mutation": brca_mutation,
+            "Overall Health": overall_health,
+            "Genomic Recurrence Score": genomic_score,
+            "Treatment": treatment,
+        }
+    )
+
+
+# Main Streamlit App
+st.title("Synthetic Data Generator")
+num_patients = st.slider("Number of Patients to Generate", 10, 1000, 100)
+if st.button("Generate Data"):
+    breast_cancer_data = generate_breast_cancer_data(num_patients)
+    st.write("Breast Cancer Data")
+    st.dataframe(breast_cancer_data)
+    st.download_button(
+        "Download Breast Cancer Data",
+        data=breast_cancer_data.to_csv(index=False),
+        file_name="breast_cancer_data.csv",
+        mime="text/csv",
+    )