OECD Pensions Explorer

• Solo project, built within a hackathon deadline.
• The data is a curated snapshot of OECD sources; it does not update live.

• Brings pension indicators that usually live in separate reports into one comparable interface.
• Four views, each answering a distinct question: generosity, cost, retirement age, and old-age poverty.
• Reproducible dataset harmonized from three OECD sources (Pensions at a Glance, SOCX, LFS).

• Start from the policy question, not the indicator at hand: each chart exists to show one concrete trade-off.
• Harmonize several official sources into a single reproducible dataset before visualizing.
• Pick the chart type by the question —choropleth, line, scatter, radar—, not for visual variety.

Pension debates usually lean on isolated metrics: a spending figure here, a replacement rate there, a retirement age in yet another report.

That fragmentation hides the core trade-off: a system can be generous, fiscally sustainable, or equitable in old age, but rarely all three at once.

The explorer puts those indicators on a single surface so that comparing countries no longer means jumping between PDFs.

• Design each view around a policy question, not around an available table.
• Harmonize Pensions at a Glance, SOCX, and LFS into a reproducible dataset instead of hard-coding figures.
• Map the chart to the question: choropleth for generosity, lines for cost and retirement age, scatter for poverty, radar for country comparison.
• Keep the dataset under the contest cap (75 MB) so the app loads fast.

• The three OECD sources are cleaned and merged in pandas into a single dataset.
• Dash orchestrates the callbacks; Plotly provides default interaction (hover, zoom, selection).
• It deploys to plotly.app as a public app, with no backend of its own to maintain.

• Won the Chief Viber prize, the top award at the Plotly Analytics Vibe-a-Thon (October 2025).
• Plotly selected it for its official finance examples: external editorial validation, not a self-assessment.
• It stands as a public, verifiable piece —live demo— and as a trade-off narrative reusable across other policy domains.

• Project on Devpost
• Contest: Plotly Analytics Vibe-a-Thon