HCP Segmentation Dashboard

Executive Summary

Real KPIs from 20,931 HCPs across 86-week longitudinal panel

Total HCPs

86-week longitudinal panel

Labeled Cohort

56.9% of total market

Unlabeled Pool

Pending classification

Feature Columns

7 feature blocks engineered

Deep Learning Recall

On minority class (SEG_C)

Data Source: All metrics are computed from hcp_analysis_clean.parquet (20,931 HCPs × 191 columns). KPIs reflect real aggregated prescribing, engagement, and market share data across the 86-week observation window.

Population Distribution

Labeled Target Breakdown

Key Metrics by Segment (Labeled Cohort)

Segment Deep-Dive

Behavioral profiling from actual 191-column feature engineering

Segment A — Traditional

Lowest prescribing volume (0.17 UC TRx/wk). Only 3.8% show Pfizer growth. Highest rep effort per Rx (0.94 details/TRx). Status-quo prescribers resistant to new therapies.

Population6,406 (53.8%)

UC TRx/week0.171

Pfizer Share of UC0.36%

Active Weeks46.1%

Details per Rx0.94

Low Priority — Baseline

Segment B — Relationship

Highest Pfizer growth signal (9.6% growing). Most efficient rep conversion (0.44 details/TRx). Broadest promo engagement (2.56 channels). The primary commercial target.

Population3,349 (28.2%)

UC TRx/week0.517

Pfizer Share of UC0.48%

Active Weeks73.3%

Details per Rx0.44

High Priority — Target

Segment C — Didactic

Highest UC volume (0.71 TRx/wk) and strongest biologic loyalty (11.3% IL-23 share). Most efficient rep relationship (0.38 details/TRx). Protocol-driven, evidence-based HCPs.

Population2,144 (18.0%)

UC TRx/week0.711

Pfizer Share of UC0.31%

Active Weeks76.8%

Details per Rx0.38

Upsell Target — Deep Learning Focus

UC TRx Volume by Segment

Medication Mix by Segment

HCP Longitudinal Journeys

Visualizing the 86-week sequential data fed into the tensors

Dr. James Chen

San Francisco, CA — Relationship-Centric

Segment B

Avg Weekly TRx Volume

14.2

86-Week Interactions

High

Dr. Sarah Williams

Chicago, IL — Didactic / Cautious

Segment C

Avg Weekly TRx Volume

6.8

86-Week Interactions

Moderate

Dr. Maria Lopez

Houston, TX — Traditional Prescriber

Segment A

Avg Weekly TRx Volume

12.0

86-Week Interactions

Low

Brand1 Adoption & Trajectory

Adoption funnel, growth signals, and recency trends from real data

91.4%

Never Tried Brand1

5.0%

Currently Active

2.6%

Trialed Then Lapsed

Adoption Funnel by Segment (% of each segment)

Adoption Funnel by Segment (Absolute Count)

Growth Signals: Who Is Accelerating?

Pfizer TRx: 86-Week Average vs. Recent 8 Weeks

Competitive Intelligence

Pfizer vs Brand2 market dynamics across segments

3.90×

Brand2/Pfizer — SEG_A

4.43×

Brand2/Pfizer — SEG_B

4.29×

Brand2/Pfizer — SEG_C

Competitive Pressure: Brand2 outprescribes Pfizer by 3.9–4.4× across all segments. The gap is widest in SEG_B (4.43×), the very segment with the highest Pfizer growth trajectory — indicating an active battleground for share.

Pfizer vs Brand2: Market Share of UC (%) by Segment

Brand2/Pfizer Prescribing Ratio by Segment

UC TRx Volume vs Pfizer Market Share (Labeled HCPs)

Rep Engagement ROI

Measuring commercial efficiency: visits, channels, and prescribing impact

0.94

Details/Rx — SEG_A

0.44

Details/Rx — SEG_B

0.38

Details/Rx — SEG_C

2.1×

SEG_B is 2.1× more efficient than SEG_A

Engagement Metrics by Segment

Rep Visits vs Pfizer Prescribing (Scatter)

Unlabeled HCP Opportunity

Prioritizing 633 unclassified HCPs for commercial outreach

Tier 1 — Immediate

Score ≥ 0.60. Highest prescribing + growth signals.

Tier 2 — Validate

Score 0.35–0.60. Moderate opportunity, needs validation.

580

Tier 3 — Monitor

Score < 0.35. Low activity, monitor for emergence.

Coverage Gap: 347 of 633 unlabeled HCPs (54.8%) have zero rep visits. Among Tier 1 (high-opportunity) HCPs, many prescribe actively but have never been contacted by a sales representative.

Live Model Prediction: Segment Classification

Test the live Hugging Face model (SEG_A vs SEG_BC) with sample HCP data.

Opportunity Score Distribution (633 Unlabeled HCPs)

Click a red point to identify the HCP below ↓