AUDIENCE & BUSINESS
Houston Grand Opera
Understanding when opera audiences become donors.
A Rice D2K collaboration with Houston Grand Opera: connecting audience profiles, donation timing, and fundraising decisions through survival analysis.
View presentation poster- Customers before cleaning & filtering
- ~13,000
- First-donation RSF · Evaluation C-index
- 0.733
The source data covered approximately 13,000 customers across survey, ticketing, donation, and marketing records. Unusable or unsuitable records were excluded during preparation; the experiments below use their eligible subsets, not all 13,000 customers. The analysis used 46 features, including age and income.
THE DECISION
Who should the team understand and engage next?
I worked across data integration, EDA, survival modeling, evaluation, and recommendations in the Rice–HGO team, and presented the findings through a poster and live Q&A.
Data Sources
Survey · Ticketing · Donations · Marketing
Features
Age · Income · Attendance · Subscription
Analysis Methods
Kaplan–Meier · Cox · Random Survival Forest
From Customer Profile to Outreach Plan
A Profile to Explore
- Age55+
- EducationBachelor’s or higher
- Household Income$125k+ / year
- LocationHouston
- RelationshipSubscriber
- EngagementHigh satisfaction & recommendation intent
A group-level profile to guide outreach exploration, not a description of every donor or an individual eligibility rule.
Recommended Actions
- Frequent Attendance
Give repeat visitors relevant follow-up and consider annual visit frequency when planning outreach.
- Subscription Relationships
Build on existing subscriber relationships and strengthen subscriber benefits.
- Strong Recommendation Intent
Support engaged customers with a good experience and opportunities to recommend HGO to others.
NPS (Net Promoter Score) summarizes how willing customers are to recommend an organization. Here, recommendation responses help describe engagement; they are distinct from overall satisfaction.
These are proposed actions informed by the analysis; resulting fundraising gains were not measured in this project.
Where Are the Customers?
Mapped customer ZIP codes to explore their geographic distribution across the United States. The heatmap helps describe audience reach; warmer areas indicate stronger concentrations in the displayed data, not higher donation rates.
When Might a Customer Donate?
Survival analysis studies how long it takes for an event to happen. Here, the event is a first or second donation. Cox regression gives an interpretable statistical model; Random Survival Forest (RSF) combines decision trees to capture more complex patterns.
| Task and Model | Training C-index | Evaluation C-index |
|---|---|---|
| First donation · Cox, full feature set | 0.731 | 0.717 |
| First donation · Random Survival Forest | 0.800 | 0.733 |
| Second donation · Cox, before feature selection | 0.603 | 0.581 |
Saved notebook outputs, rounded to three decimals, from 80/20 splits. These are exploratory results: evaluation data were also consulted during feature and parameter selection. They are not a fresh, untouched final test. The poster’s second-donation score of 0.61 is from a full-data fit; feature-removal experiments reached about 0.625 on the reused evaluation split.
Reading the Results
C-index · Which Customer Donates Earlier?
Measures how well the model orders donation times among customer pairs that can be compared. 0.5 is chance-level ordering; 1.0 is perfect ordering. A score of 0.733 is roughly 73 out of 100 comparable pairs ordered correctly, with ties receiving partial credit. It does not mean 73.3% of customers will donate.
Survival Probability · Not Yet Donated
In these plots, “survival” means the donation has not happened yet. A value of 0.7 at year 2 means an estimated 70% chance of no donation by then; 1 − 0.7 = 30% is the estimated chance of a donation by that time. A faster fall indicates earlier predicted giving.
These curves illustrate model predictions, rather than observed conversion rates.
Interactive Demo · Rule-Based Simulation
From Customer Features to Decision Support
Explore a customer scenario, then filter a 500-customer simulated audience. This demo illustrates how customer analysis can become an input-and-output workflow and a BI view for fundraising discussions.
The original university–industry project uses private data and models that are not served here. This portfolio demo uses 500 synthetic customers and illustrative rules informed by the project discussion: little attendance and greater distance reduce simulated donation propensity. Rule strengths are set for demonstration, not fitted coefficients or outputs from the trained model.
Try a Customer Scenario
This scenario is a separate rule-based example, not a filter on the board below. Distance is an illustrative input, not a verified distance coefficient from the trained model.
Explore the 500-Customer Simulation
Donation Participation by Attendance
What to Investigate Next
Metric Definitions
- Audience
- Customers in the selected region and subscription group. This synthetic cohort is fixed at the start of the demo year.
- Donation Participation
- Customers with at least one donation in the selected window ÷ customers in the selected group. This is not an outreach conversion rate.
- Repeat-Donation Share
- Customers with two or more donations in the window ÷ customers with at least one donation in that same window. This is not next-period retention.
- Donation Amount
- Total value of donations within the selected window, in simulated US dollars. Attendance bands also use that window.
