How a Philanthropic Fund Brought Its Funding Data Into One View
A leading philanthropic fund advancing human rights and economic justice runs programmes across several regions and focus areas. Its funding reporting was split across international, US and programme-level views, which made overall performance hard to see. OnPoint Insights built an interactive R Shiny application that brings allocations, approvals and remaining budgets into one place, by country, state and programme.
3 levels of view
Funding analysed by country, state and programme, in a single interactive application.
Questions answered in the review
Filters and export-ready views mean a stakeholder question is answered during the allocation review, not in a follow-up pack afterwards.
Open source, handed over
Documented, modular R code the fund’s own team can extend and maintain, with no license attached.
The problem
The money was tracked. The picture was not.
For a funder, the hard question is rarely how much has been granted. It is where the next allocation does the most good, and that answer only appears when approvals, remaining budgets and year-on-year movement can be seen next to each other.
Our client funds programmes across multiple regions and focus areas. The data existed and the reporting worked, but it was split by funding stream, which made it difficult to analyse allocations across geographies, programmes and approval stages in one place.
What was getting in the way
- Reporting was fragmented across international, US and programme-level funding, so overall financial performance was never visible in one view.
- Grants, allocations, approval rates, remaining budgets and year-on-year trends were difficult to see together.
- There was no filterable, easy-to-navigate interface for tabular reporting or stakeholder review.
- Any solution had to be open source, so the team could maintain and extend it internally.
Our approach
Make the tables do the work
What we found
The fund did not need another wall of charts. It needed its funding tables to answer questions: what has been approved, what is left to allocate, and how that compares with last year. Keeping the format tabular was the point, because that is how allocation decisions are reviewed.
What we did
Data prepared properly, in R. Funding data was cleaned, transformed and structured using dplyr, tidyr and lubridate, so every figure in the application comes from one consistent source.
Three hierarchical views. Tables by country, state and programme show approved and pending amounts, remaining budgets, approval percentages and year-on-year trends as sparklines.
Filters and visual cues built for reviewers. Region, fiscal year and programme type filters, with bar graphs and conditional formatting, so the outliers stand out without reading every row.
Reporting that leaves the screen. Exports to CSV, Excel and PDF, plus print-ready views, so the same numbers go straight into internal reporting and stakeholder packs.
The choice we made
We built the application in open-source R and Shiny and handed over documented, modular code rather than a licensed platform. The fund’s own team can point it at new datasets and programmes without coming back to us, and without a licence line in the budget.
-
R Shiny -
R
The results
What changed for the team
Funding analysis stopped being an exercise in assembling reports and became something the team can do live.
Funding is visible across every dimension
Allocations can be explored by country, state and programme, with approvals, remaining budget and year-on-year movement in the same table rather than in three separate reports.
Questions get answered in the room
Filters and export-ready views mean a stakeholder question is answered during the review, instead of becoming a follow-up pack a week later.
The framework belongs to the client
The modular codebase can be pointed at new funding datasets or programmes, so the next piece of reporting starts from something, not from nothing.
Delivered by a Microsoft Solutions Partner for Data & AI
Facing a similar challenge?
Talk to our data and AI team about where you are today and what good looks like. 30 minutes, no obligation.
Book a consultation
More Case Studies
Next Case Study
How a Preventive Healthcare Provider Turned Diagnostic Data into Early Warnings
How an HVACR Retailer Rebuilt Product Recommendations with Machine Learning
