Atlas of Astronaut Collaboration and Foresight

Inspiration

Space is often told as rockets and heroes. We asked a social-science question instead: what does the structure of human-spaceflight cooperation say about world politics—and do inflation shocks (CPI) actually change how often humans launch? Our hunch: launch tempo is governed by institutions, vehicles, and program eras, not by yearly price swings. That became the spine of our project.

What it does

Atlas of Astronaut Collaboration and Foresight turns mission logs into a country-to-country network by year and layers it with context:

Annual cooperation graphs with hubs, communities, and edge weights (co-missions).
Small multiples for the US, Russia, and Japan: launch frequency vs CPI with event markers (1986 Challenger, 2003 Columbia, 2011 Shuttle retirement, 2020 Commercial Crew).
Metrics for analysis and policy: betweenness (“brokers”), a Free-Ride Index (share of a country’s missions that include a hub), inequality of access (Gini of seats), and resilience tests (what breaks if a hub exits).

How we built it

Data: CSIS International Astronaut Database (+ decade Wikipedia tables where needed) → astronauts, nationalities, missions. World Bank WDI → annual CPI (US/RU/JP).
Engineering: cleaned and standardized countries (ISO, USSR→Russia, etc.), parsed dates, and built a dyad–year edge list (mission-level co-crew → yearly aggregation).
Network science: NetworkX for degree/weighted degree, betweenness, community detection; robustness via “remove-a-hub” experiments.
Quant & viz: pandas/statsmodels for event windows and simple count models; matplotlib/plotly for network snapshots and CPI overlays.
Reproducibility: one-click notebooks, versioned CSVs, and a data dictionary.

Challenges we ran into

Heterogeneous tables (merged cells, flags instead of country text), dual nationalities, and historic states mapping.
Deciding on fair weights (missions vs seats vs crew-days) and handling multi-country crews.
Aligning social indicators to mission calendars (lags, program eras) so tests aren’t spurious.
Keeping the CPI story honest—CPI ≠ “welfare,” and budgets are multi-year/planned.

Accomplishments that we’re proud of

A clean, reproducible country–country–year cooperation dataset.
Clear evidence that structural eras and vehicles explain launch tempo better than CPI.
Intuitive metrics judges remember: Free-Ride Index, access inequality (Lorenz/Gini), and resilience under hub removal.
Polished visuals that make the diplomacy behind space tangible.

What we learned

Human spaceflight is planned and lumpy—a supply-constrained system where certification, fleet readiness, and international governance dominate short-run prices.
Network views translate a tech domain into social structure: who brokers ties, who piggybacks, and how blocs realign.
Good social-science stories need falsification tests (here, CPI as a placebo) and transparent assumptions.

What’s next for Atlas of Astronaut Collaboration and Foresight

Add UN General Assembly voting similarity and formal alliance ties to explain which cooperation edges form (QAP/MRQAP).
Bring in real (CPI-deflated) space budgets to test the capacity/vehicle mechanism directly.
Switch outcomes from mission counts to seat-counts / crew-days, plus link-prediction for “next-decade partners.”
Ship an interactive web atlas (filter by era, simulate hub exits, export country fact sheets). One-line vision: from “who flew with whom” to a foresight tool for how cooperation shifts when politics or fleets change.