Create app.py

app.py

@@ -0,0 +1,233 @@
+import math import numpy as np import pandas as pd import gradio as gr
+
+------------------------------
+
+Utility helpers
+
+------------------------------
+
+def usd(x: float) -> str: try: return "$" + f"{x:,.0f}" except Exception: return "$0"
+
+def usd2(x: float) -> str: try: return "$" + f"{x:,.2f}" except Exception: return "$0.00"
+
+def pct(x: float) -> str: return f"{x*100:.1f}%"
+
+------------------------------
+
+Core calculator
+
+------------------------------
+
+def compute_roi( scans_per_month: int, # Clinical (detection) base_ppr: float,  # baseline positive pickup rate (as fraction) ai_ppr: float,    # with-AI positive pickup rate (as fraction) base_audit_rate: float, ai_audit_rate: float, # Operational (time) base_read_seconds: float, read_reduction_pct: float,  # as fraction radiologist_hourly_cost: float, # Financial (per-scan cost) base_cost_per_scan: float, cost_reduction_pct: float, # Financial (follow-ups) followup_price: float, followup_uplift_pct: float,  # as fraction of scans # Recalls / false positives base_recall_rate: float, ai_recall_rate: float, recall_cost_per_case: float, # Early detection value early_detect_uplift_per_1000: float,  # extra cancers detected per 1000 treatment_cost_delta_early_vs_late: float,  # savings per earlier case # Program costs ai_subscription_monthly: float, integration_overhead_monthly: float, cloud_compute_monthly: float, ): # --- Clinical metrics --- errors_reduced = max(0, scans_per_month * (base_ppr - ai_ppr))  # if AI lowers miss rate via higher pickup, set accordingly discrepant_flags = max(0, scans_per_month * (ai_audit_rate - base_audit_rate)) earlier_detections = max(0.0, scans_per_month * (early_detect_uplift_per_1000 / 1000.0)) recalls_avoided = max(0, scans_per_month * (base_recall_rate - ai_recall_rate))
+
+# --- Operational metrics ---
+hours_saved = scans_per_month * (base_read_seconds * read_reduction_pct) / 3600.0
+workload_reduction_pct = read_reduction_pct
+fte_saved = hours_saved / 160.0  # rough FTE-month heuristic
+capacity_increase_pct = (1.0 / max(1e-6, (1.0 - read_reduction_pct)) - 1.0)  # approximate capacity headroom
+
+# --- Financial metrics ---
+value_time_saved = hours_saved * radiologist_hourly_cost
+
+baseline_monthly_cost = scans_per_month * base_cost_per_scan
+new_monthly_cost = baseline_monthly_cost * (1.0 - cost_reduction_pct)
+per_scan_cost_savings = baseline_monthly_cost - new_monthly_cost
+
+addl_followups = scans_per_month * followup_uplift_pct
+addl_followup_revenue = addl_followups * followup_price
+
+recall_cost_savings = recalls_avoided * recall_cost_per_case
+early_detection_savings = earlier_detections * treatment_cost_delta_early_vs_late
+
+total_monthly_benefit = (
+    value_time_saved
+    + per_scan_cost_savings
+    + addl_followup_revenue
+    + recall_cost_savings
+    + early_detection_savings
+)
+
+total_monthly_costs = ai_subscription_monthly + integration_overhead_monthly + cloud_compute_monthly
+net_roi = total_monthly_benefit - total_monthly_costs
+roi_pct = (net_roi / total_monthly_costs) if total_monthly_costs > 0 else np.nan
+
+# One-line summary
+clinical_bullet = (
+    f"~{int(round(recalls_avoided))} recalls avoided, "
+    f"{earlier_detections:.1f} earlier cancers detected, "
+    f"{int(round(errors_reduced))} fewer missed positives"
+)
+summary = (
+    f"For your practice with {scans_per_month:,} mammography scans/month, your net benefit is "
+    f"{usd(net_roi)} per month. Clinical impact: {clinical_bullet}."
+)
+
+# Build Markdown sections
+financial_md = f"""
+
+Financial (USD/month)
+
+Metric	Value
+
+Value of radiologist time saved	{usd(value_time_saved)}
+Savings from per-scan radiologist cost	{usd(per_scan_cost_savings)}
+Additional follow-up revenue	{usd(addl_followup_revenue)}
+Savings from avoided recalls	{usd(recall_cost_savings)}
+Savings from earlier detection	{usd(early_detection_savings)}
+Total monthly benefit	{usd(total_monthly_benefit)}
+AI subscription	{usd(ai_subscription_monthly)}
+Integration / training / overhead	{usd(integration_overhead_monthly)}
+Cloud / compute	{usd(cloud_compute_monthly)}
+Total monthly costs	{usd(total_monthly_costs)}
+Net monthly ROI	{usd(net_roi)}
+ROI %	{'' if math.isnan(roi_pct) else f'{roi_pct*100:.1f}%'}
+"""	
+
+
+clinical_md = f"""
+
+Clinical (per month)
+
+Metric	Value
+
+Fewer missed positives (Δ pickup)	{int(round(errors_reduced))}
+Discrepant cases flagged by AI (audit uplift)	{int(round(discrepant_flags))}
+Earlier cancers detected	{earlier_detections:.1f}
+Recalls avoided	{int(round(recalls_avoided))}
+"""	
+
+
+operational_md = f"""
+
+Operational
+
+Metric	Value
+
+Hours of radiologist reading time saved / month	{hours_saved:.1f}
+Workload reduction	{pct(workload_reduction_pct)}
+Approx. FTE-month saved	{fte_saved:.2f}
+Effective reading capacity increase	{pct(capacity_increase_pct)}
+"""	
+
+
+# CTA
+cta_md = (
+    "**Want to see this live in your workflow?**  "
+    "[Book a demo →](https://example.com/book-demo)"
+)
+
+return summary, financial_md, clinical_md, operational_md, cta_md
+
+------------------------------
+
+UI Layout
+
+------------------------------
+
+def build_ui(): with gr.Blocks(title="Mammography AI ROI Calculator", theme=gr.themes.Soft(), css=""" .gradio-container {max-width: 1200px !important} .section-head {font-weight: 700; margin-top: .5rem} .small-note {opacity: .70; font-size: 0.9em} """) as demo: gr.Markdown("""
+
+Mammography AI ROI Calculator
+
+A clean, defensible ROI model for breast imaging AI — with Financial, Clinical, and Operational impact, all in USD. """) with gr.Row(): with gr.Column(scale=1): gr.Markdown("## Inputs") scans_per_month = gr.Number(value=7500, label="Scans per month", precision=0)
+
+with gr.Accordion("Clinical inputs", open=False):
+                base_ppr = gr.Slider(0, 1, value=0.10, step=0.001, label="Baseline positive pickup rate (fraction)")
+                ai_ppr = gr.Slider(0, 1, value=0.095, step=0.001, label="With-AI positive pickup rate (fraction)")
+                base_audit_rate = gr.Slider(0, 1, value=0.00, step=0.001, label="Baseline audit flag rate (fraction)")
+                ai_audit_rate = gr.Slider(0, 1, value=0.05, step=0.001, label="With-AI audit flag rate (fraction)")
+                base_recall_rate = gr.Slider(0, 1, value=0.028, step=0.001, label="Baseline recall rate (fraction)")
+                ai_recall_rate = gr.Slider(0, 1, value=0.025, step=0.001, label="With-AI recall rate (fraction)")
+                recall_cost_per_case = gr.Number(value=250.0, label="Cost per recall case (USD)")
+                early_detect_uplift_per_1000 = gr.Number(value=0.7, label="Earlier cancers detected (extra per 1000 scans)")
+                treatment_cost_delta_early_vs_late = gr.Number(value=15000.0, label="Treatment cost savings per earlier case (USD)")
+
+            with gr.Accordion("Operational inputs", open=False):
+                base_read_seconds = gr.Number(value=100, label="Baseline read time per scan (seconds)")
+                read_reduction_pct = gr.Slider(0, 0.8, value=0.15, step=0.005, label="Reading time reduction with AI (fraction)")
+                radiologist_hourly_cost = gr.Number(value=180, label="Radiologist hourly cost (USD)")
+
+            with gr.Accordion("Financial inputs", open=False):
+                base_cost_per_scan = gr.Number(value=15, label="Radiologist cost per scan (USD)")
+                cost_reduction_pct = gr.Slider(0, 0.8, value=0.15, step=0.005, label="Per-scan radiologist cost reduction (fraction)")
+                followup_price = gr.Number(value=200, label="Price per follow-up scan (USD)")
+                followup_uplift_pct = gr.Slider(0, 0.2, value=0.0095, step=0.0005, label="Follow-up uplift (fraction of scans)")
+
+            with gr.Accordion("Program costs", open=True):
+                ai_subscription_monthly = gr.Number(value=0, label="AI subscription (USD/month)")
+                integration_overhead_monthly = gr.Number(value=0, label="Integration / training / overhead (USD/month)")
+                cloud_compute_monthly = gr.Number(value=0, label="Cloud / compute (USD/month)")
+
+            calc_btn = gr.Button("Calculate ROI", variant="primary")
+
+        with gr.Column(scale=1):
+            summary = gr.Markdown("", elem_id="summary")
+            with gr.Tab("Financial"):
+                financial_md = gr.Markdown()
+            with gr.Tab("Clinical"):
+                clinical_md = gr.Markdown()
+            with gr.Tab("Operational"):
+                operational_md = gr.Markdown()
+            gr.Markdown("---")
+            cta_md = gr.Markdown()
+
+    calc_btn.click(
+        fn=compute_roi,
+        inputs=[
+            scans_per_month,
+            base_ppr,
+            ai_ppr,
+            base_audit_rate,
+            ai_audit_rate,
+            base_read_seconds,
+            read_reduction_pct,
+            radiologist_hourly_cost,
+            base_cost_per_scan,
+            cost_reduction_pct,
+            followup_price,
+            followup_uplift_pct,
+            base_recall_rate,
+            ai_recall_rate,
+            recall_cost_per_case,
+            early_detect_uplift_per_1000,
+            treatment_cost_delta_early_vs_late,
+            ai_subscription_monthly,
+            integration_overhead_monthly,
+            cloud_compute_monthly,
+        ],
+        outputs=[summary, financial_md, clinical_md, operational_md, cta_md],
+    )
+
+    # Auto-run with defaults
+    demo.load(
+        fn=compute_roi,
+        inputs=[
+            scans_per_month,
+            base_ppr,
+            ai_ppr,
+            base_audit_rate,
+            ai_audit_rate,
+            base_read_seconds,
+            read_reduction_pct,
+            radiologist_hourly_cost,
+            base_cost_per_scan,
+            cost_reduction_pct,
+            followup_price,
+            followup_uplift_pct,
+            base_recall_rate,
+            ai_recall_rate,
+            recall_cost_per_case,
+            early_detect_uplift_per_1000,
+            treatment_cost_delta_early_vs_late,
+            ai_subscription_monthly,
+            integration_overhead_monthly,
+            cloud_compute_monthly,
+        ],
+        outputs=[summary, financial_md, clinical_md, operational_md, cta_md],
+    )
+
+return demo
+
+def main(): return build_ui()
+
+if name == "main": app = build_ui() app.launch()
+