HomeWorkServicesAboutBlogContact
Start a Project
Back to Work
Dashboard·Data Intelligence·Global Philanthropy

Gates Foundation

Turning billions of data points into decisions that matter.

Client
Bill & Melinda Gates Foundation
Year
2024
Scope
12 Weeks · Independent Designer
Gates Foundation · Program Intelligence
142
Programs Active
$6.2B
Funding Deployed
58
Countries
Regional Impact
Health
Education
Agriculture
40%
Faster report interpretation
60%
Improvement in metric clarity
75%
Rise in non-technical adoption
5s
Design standard, insight visible per view
01·Context

Data abundant. Insight scarce.

The Bill & Melinda Gates Foundation operates at a scale few organisations can match, spanning global health, education, poverty alleviation, and agricultural innovation across continents. Their programs generate thousands of datasets collected from partner organisations, research teams, and field operations. Every metric is carefully captured. Every dataset is meticulously maintained.

The problem wasn't the data. It was what happened to it afterwards. Presented through dense spreadsheets, disconnected dashboards, and static presentations, the data made it nearly impossible to extract meaning quickly, particularly for program directors making decisions across multiple regions simultaneously.

Dense
spreadsheets and static decks as primary outputs
Fragmented
dashboards across teams with no shared visual language
Hours
of manual analysis before each program review

The foundation recognized that meaningful design could be the bridge between raw data and real-world impact. Clarity was not an aesthetic goal, it was a strategic one.

Project brief, Gates Foundation
02·Role

Independent designer, visualization strategy to delivery.

Brought in as an independent designer, the remit covered the full visualization and dashboard design initiative, from understanding decision workflows through to chart selection, hierarchy design, and handoff to the foundation's internal data engineering team.

Understanding data workflows and user decision patterns through director interviews
Developing visual design principles for the entire dashboard system
Selecting appropriate chart types mapped to each data structure and decision type
Designing modular dashboard layouts tailored for each program area
Building the component library with chart cards, filters, and interaction specs
Ensuring clarity, accessibility, and brand consistency across all touchpoints
03·The Challenge

Four data problems, none of them technically about data.

The Gates Foundation's challenge was not a data quality problem. The data was excellent. It was a data communication problem, and communication failures at this scale translate directly into slower decisions, misallocated resources, and missed opportunities for impact.

Cognitive Overload

Directors interpreting large volumes of tabular data simultaneously
Key insights buried under numbers with no visual guidance
Extended analysis sessions producing decision fatigue before action

Fragmented Data Sources

Health, education, and agriculture teams using separate dashboards
Inconsistent metrics and duplicate visualisations across departments
No shared visual language for cross-team data discussions

Lack of Visual Hierarchy

Generic templates with no consistent design logic applied
Primary KPIs visually indistinguishable from supporting metrics
Users unable to orient themselves quickly in any given view

Misaligned Chart Selection

Pie charts used for time-series data, causing misreads
Clustered bars applied to comparative distributions they couldn't communicate
Wrong chart types eroding trust in the visualization system itself
04·Data Taxonomy

Chart selection starts with categorising the data first.

The first step wasn't opening Figma, it was interviewing program directors to understand how they made decisions and what questions they needed answered in real time. Their answers revealed clear patterns: regional comparisons, funding utilisation, outcome correlations, and implementation bottlenecks. Each pattern mapped to a data type, and each data type to a visual form.

Data type
Decision it informs
Chart selected
Comparative
Program performance across countries or intervention types
Horizontal bar charts, clean, scannable, directly comparable
Time-Series
Longitudinal outcome tracking and funding utilisation over time
Line charts with milestone markers and confidence intervals
Proportional
Budget allocation ratios and beneficiary segmentation
Donut charts and stacked bars, not pie charts
Correlation
Relationship between funding levels and outcomes achieved
Scatter plots with trend lines and outlier annotation
Geographic
Regional impact distribution and country-level performance
Choropleth maps with quantitative colour scales
05·Dashboard Hierarchy

Three layers of detail. One navigable system.

The dashboards were structured around three levels of interaction, each designed for a different decision-making context. Directors needed to move fluidly between macro portfolio views and micro country-level details without changing tools, tabs, or mental models.

Executive Overview

Portfolio-wide KPIs
Cross-program funding summary
At-risk program flags
Board-ready visual summaries
5-second rule enforced

Program-Level Analysis

Performance by focus area
Funding vs outcome correlation
Trend deviations and anomalies
Partner delivery status
Cross-program comparison

Regional Insights

Country-specific breakdowns
Drill-down to district level
Demographic filters
Field-level operational data
Exportable country reports
06·Design Principles

Four rules that governed every element on every screen.

These weren't aesthetic guidelines. They were functional constraints, each one traceable to a specific failure mode observed in the existing dashboards and designed to eliminate it.

The Five-Second Rule

Every dashboard view designed to communicate its most critical insight within five seconds. Primary KPIs at top-left, the quadrant the eye naturally lands on first.

Data-Ink Ratio Optimisation

Inspired by Tufte, every gridline, border, and decorative element that didn't serve an analytical purpose was removed. The visual system earned its complexity.

Layout Architecture

Critical KPIs top-left, comparative charts top-right, detail tables below. Consistent white space prevented crowding. Users navigated by instinct, not by reading.

Strategic Colour Use

Minimal palette optimised for data differentiation, not aesthetics. Shape, density, and typography carried hierarchy alongside colour. Red reserved strictly for alert states.

07·Visualization Modules

Four program areas. Four purpose-built solutions.

Each module was designed around the specific analytical questions its program area needed to answer. The chart types differ because the data structures and decision contexts differ, but the visual grammar, spacing, and hierarchy principles remain consistent throughout.

01

Health Impact Dashboards

Progressive line charts tracking vaccination rates and healthcare access over time, with geographic overlays providing simultaneous spatial context. Demographic and regional filters for specific population cohorts.

Line charts + Choropleth overlays + Demographic filters
02

Education Outcomes Visualisation

Multi-metric panels combining line and bar charts for literacy improvements, enrollment rates, and teacher-to-student ratios. Interactive correlation filters to test funding-to-outcome relationships directly.

Small multiples + Correlation scatter + Funding overlay
03

Agriculture & Livelihood Modules

Bubble maps combining data density and performance indicators into a single spatial view. Hover states revealing live data points, intuitive for non-technical program managers.

Bubble maps + Regional density + Hover-to-detail
04

Funding Utilisation & Budget Tracking

Sankey diagrams representing the complete fund flow, from donors through programs to beneficiaries. Increased transparency, improved stakeholder trust, and made donor reporting a visual conversation.

Sankey flow + Budget vs actuals + Stakeholder audit view
08·Accessibility & Inclusivity

Designed for everyone, analysts and non-analysts alike.

The biggest failure of the previous system wasn't data coverage, it was that only data analysts could actually use the dashboards effectively. Accessibility work here was about expanding who could engage with the data.

Colorblind-Safe Palettes

All charts used distinguishable colours with pattern overlays as alternatives. No information encoded in colour alone, shape, density, and position were always secondary carriers.

Tooltips & Labels for All Data Points

Every data point had a tooltip providing exact values and contextual labels. Charts never required users to estimate, the precise value was always one hover away.

Keyboard Navigation

All dashboard interactions, chart selection, filter application, drill-down navigation, fully keyboard-navigable, meeting WCAG 2.1 AA compliance standards.

Mobile-Friendly Layouts

Responsive layouts adapted across desktop, tablet, and mobile for program directors needing quick-reference access. Priority information reordered vertically for smaller screens.

09·Impact

Measurable results across every layer of the organisation.

40%
Reduction in time spent interpreting reports, hours of manual analysis now at a glance
60%
Improvement in multi-dimensional metric comprehension among program heads
75%
Increase in dashboard use beyond analysts, program managers and field leads now self-serve
Unified
Health, education, and agriculture teams now discuss data using one shared visual language

Each design decision, from chart type to layout, was guided by one principle: clarity drives action. The dashboards did not just visualize impact. They enabled it.

Project summary, Gates Foundation
10·Learnings

What designing for global development intelligence demands.

01

Context is everything in chart selection

A visualisation is only as powerful as the decision it serves. Understanding what questions program directors needed answered, not what data existed, was the only reliable way to choose the right chart.

02

Simplicity is strategic, not reductive

Removing unnecessary complexity didn't simplify the data, it clarified it. The challenge was to communicate depth without overwhelming. Every element removed was a potential distraction eliminated.

03

Design and data are equal partners

Visualization succeeds only when design and data structure evolve together. Working closely with the foundation's analytics team ensured every visual choice was grounded in validated data structures.

04

Accessibility multiplies impact

Designing for accessibility didn't constrain the system, it expanded its reach. The improvements in contrast, labelling, and legibility made the dashboards more effective for everyone.

The Gates Foundation had rich, meticulously collected data across health, education, agriculture, and development. What they lacked was a visual system that could make that data readable, comparable, and actionable for the program directors who needed to act on it fastest.

More Work

Like what you see?

Let's build something worth a case study.