In collaboration with non-profit organizations, I’ve developed predictive models and interactive dashboards to support national campaign strategy and local outreach optimization. This project spans multiple phases and use cases across voter engagement analytics.

Below are selected highlights from this ongoing initiative.

---

Selected Projects

• Segmentation via Clustering

  Applied unsupervised learning (HDBSCAN) to identify voter segments with varying degrees of partisanship and persuadability. These insights inform message tailoring and targeting strategies. (To learn more, see Clustering Analysis)

• Geospatial Prioritization

  Built models to rank neighbourhoods based on their outreach responsiveness using demographic, behavioral, and electoral data. Deployed insights via a stakeholder-facing dashboard for field teams and campaign strategists. (See the Impact Voters Dashboard for a live example)