WE ARE THE MOST CONNECTED PEOPLE
WHO EVER LIVED ALONE.
One in four adults globally reports frequent loneliness (Meta-Gallup 2023). The average American's count of close confidants has fallen from 2.94 in 1985 to 1.95 in 2022 (GSS). US 15-24-year-olds spend roughly a third as much in-person time with friends as they did in 2003 (ATUS). The UK has lost ~45% of its pubs since 1980 (BBPA); the US has lost about half its fraternal-order membership over the same window (Putnam, updated). Connection is everywhere; company isn't.
Six figures behind the composite.
One 0-100 score.
Heuristic seed snapshot. Isolation stress sits firmly in the 'Severe' band. One in four adults globally reports frequent loneliness. Single-occupancy households are now ~28% of US households (~40% in some major European cities). The average American's close-confidants count has roughly halved since 1985. Adolescent in-person hangout time has collapsed; third-place infrastructure has been hollowing out for decades. The US Surgeon General, the UK government, and the Japanese government have all formally designated loneliness a public-health issue.
Over time.
Confidants, teen in-person time, single-person households, and UK pubs.
- Avg US close confidants2.94 in 1985 → 1.95 in 2022
- Teen daily in-person time (min)150 min in 2003 → 40 min in 2023
- US single-person households (%)13% in 1960 → 28% in 2023
- UK open pubs (thousands)69K in 1980 → 38K in 2024
Loneliness is no longer an old-person problem.
Pew + Meta-Gallup data show the steepest band on the curve is 16-24-year-olds. The 'shut-in elder' picture that dominated the 1980s public-health imagination has been overtaken by an even larger isolation problem among the youngest adult cohort. The U-shape persists, but the left peak is now higher than the right.
Where citizens feel most alone.
Share of adults saying they feel 'very' or 'fairly often' lonely. Türkiye and South Korea lead; the Netherlands and Sweden show that wealth and urbanisation alone don't cause it.
- 🇹🇷TürkiyeMENA33%
- 🇰🇷South KoreaAsia31%
- 🇺🇸United StatesAmericas28%
- 🇮🇹ItalyEurope27%
- 🇧🇷BrazilAmericas27%
- 🇲🇽MexicoAmericas26%
- 🇫🇷FranceEurope25%
- 🇬🇧United KingdomEurope24%
- 🇯🇵JapanAsia24%
- 🇩🇪GermanyEurope21%
- 🇮🇳IndiaAsia18%
- 🇸🇪SwedenEurope16%
- 🇳🇱NetherlandsEurope14%
What the score is measuring.
Several traditions reading the same data.
Why is the most-connected generation also the most alone?
Loneliness shows mortality-risk magnitudes comparable to smoking 15 cigarettes per day. Treating it as a clinical and policy issue, like obesity or sleep loss, is overdue. The US Surgeon General, the UK government, and Japan all now formally treat it as such. Interventions exist; the political will to scale them does not.
“We are facing an epidemic of loneliness and isolation.”
Civic infrastructure (clubs, churches, leagues, fraternal orders, PTAs) was a load-bearing element of mid-20th-century life that has been allowed to decay. The decay is faster in lower-income, lower-education communities. The civic recession and the political recession have the same shape because they have the same cause.
Smartphones + social platforms substitute screen-mediated weak ties for in-person strong ties. The substitution accelerated after 2012 and the adolescent measures suggest it is approaching a floor. The substitution is not necessarily destructive at the margin, but it has happened at a scale and pace incompatible with the wellbeing patterns the data shows.
Many Indigenous frameworks treat isolation as a symptom of severed relationships: to land, to kin, to ancestors, to the more-than-human world. The Western individual-focused frame asks 'how do we cure isolated individuals'; this frame asks 'how do we repair the relational fabric they fell out of'.
Japan and South Korea show how collectivist family structures do not automatically protect against isolation. Even where the family unit remains strong on paper, the elder hikikomori and shut-in phenomena have grown for decades. The lesson: cultural collectivism is not interchangeable with active community participation.
Where churches and equivalent congregations have closed, the social infrastructure they provided (childcare, mutual aid, weekly contact obligations across ages) has not been replaced. Re-investment in congregations and equivalent structures is treated as a primary intervention by traditions that already practice it.
Suburban-sprawl-by-default zoning destroyed the proximity that made third places economically viable. The car commute is structurally isolating. Walkable cities, mixed-use blocks, density that puts you in casual contact with neighbours: built-environment fixes deliver more loneliness reduction per dollar than most therapeutic ones.
Of the interventions, which reduce isolation?
GPs refer patients to community choirs, walking groups, library programmes. UK NHS has scaled this; preliminary evidence shows reduced loneliness scores and reduced primary-care reattendance. Replicable elsewhere with modest investment.
Restoring adolescent in-person time requires removing the device that occupies it. Phone-free schools, no smartphones before high school, age limits on social media. Pre-2010 free-range play was a relational infrastructure that the current generation has not been given.
Subsidise pubs, cafés, libraries, community centres, places of worship as civic infrastructure. The economic arguments for this are weak in market terms; the social-return-on-investment evidence is strong.
Zone for proximity. The single largest demographic determinant of loneliness scores is whether daily errands require a car. Re-permit accessory dwellings, mixed-use ground floors, narrower streets, shorter blocks.
From the practitioners: the highest-leverage intervention is weekly volunteer-driven contact for socially-isolated older adults. Programs like Re-engage (UK), AARP Connect2Affect (US), Tonari Kinjo (Japan) deliver measurable wellbeing improvements at low cost.
Where loneliness runs high, depression does too.
Loneliness and clinical depression / anxiety move together across countries. Türkiye and South Korea anchor the high corner; the Netherlands and Sweden anchor the low corner. The relationship is causal in both directions: loneliness raises depression risk, and depression in turn intensifies isolation.
What is being tried, and where.
Six interventions, at three policy levels.
Government recognition (loneliness ministers), clinical-pathway changes (social prescribing), regulatory limits (under-16 social-media bans), third-place subsidy. Each carries different evidence weight; each is at a different scale.
- AUSTRALIA2025
Under-16 social media ban
First national-level statutory age limit on major social-media platforms. Targets the documented adolescent isolation curve.
EVIDENCEToo early for clean evidence; precedent triggered similar EU member-state proposals. - UNITED STATES2023
Surgeon General Advisory
Our Epidemic of Loneliness and Isolation formal advisory. Six-pillar national strategy. No funded national programme attached.
EVIDENCEAdvisory only; municipal pickups (Boston, LA county) show some traction. - JAPAN2021
Loneliness office under Cabinet
Created in response to pandemic-period suicide increases, particularly among women. Coordinates municipal initiatives, hotlines, hikikomori support.
EVIDENCEImplementation early; cross-ministry coordination judged effective on hotline reach. - SWEDEN2020
'Folkets Hus' renewal funding
State-municipal funding to renew the historic 'People's Houses' community-centre network: a classic Nordic third-place model.
EVIDENCELocal rather than national-survey; visitor-count growth where renewal completed. - UNITED KINGDOM · NHS2019
Social prescribing at scale
GPs refer patients to community choirs, walking groups, library programmes, gardening, befriending. NHS Universal Personalised Care framework.
EVIDENCEStrong. Reduced primary-care reattendance and improved loneliness self-report in trials. Now ~1M referrals per year. - UNITED KINGDOM2018
World's first loneliness minister
Tracey Crouch appointed. Built into the Office for Civil Society, ~£23M strategy. Modest measurable effects but precedent-setting at scale.
EVIDENCEMixed. Funding modest relative to need. Strategy framework now adopted by other countries.
Sources, weights, and code are open.
Where every number comes from
The composite index is computed from the signals listed on this page, each backed by one or more named sources. Where the source publishes a public dataset or feed it is linked below; where a signal involves qualitative judgement, the LLM-assisted pass is explicitly marked on the signal card.
- ·AARP Loneliness in Older Adults
- ·ATUS (US)
- ·British Beer & Pub Assoc.
- ·CDC YRBSS
- ·EU Time Use Survey
- ·European Social Survey
- ·Eurostat
- ·Jonathan Haidt research
- ·Meta-Gallup State of Social Connections
- ·Monitoring the Future
- ·OECD
- ·OECD Better Life Index
- ·Pew Research
- ·Pew civic engagement
- ·Pew religion
- ·Putnam Bowling Alone series
- ·UK Campaign to End Loneliness
- ·UK Minister for Loneliness; US Surgeon General Advisory; Japan loneliness office
- ·UN DESA Family Database
- ·US Census ACS
- ·US General Social Survey (GSS)
- ·WHO
- ·World Values Survey
Everything is versioned
- → Every hourly snapshot is committed to git with a message naming the signals that moved.
- → A daily snapshot is archived to
data/history-current/for the calibration log. - → Raw scraped article lists are written to
data/raw/so a score is reproducible from its input bundle. - → Signal definitions, weights, and seeded scores all live in plain JSON or TypeScript; anyone can open a PR challenging a value and explain why.
How this pillar is scored.
Methodology & limits
Ten signals, weighted into a 0-100 score. Half of them are individual-experience measures (loneliness rates, close-confidant counts, sleep displacement, in-person time). The other half are structural (single-occupancy households, third-place decline, civic-organisation participation, formal policy recognition).
Adolescent in-person time gets a heavy weight because the curve is the steepest cohort-level behavioural change of the last twenty years. The knock-on effects in mental health, polarisation, and educational outcomes are well documented.
Two adjacent pillars share data with this one: mental health (the same loneliness and sleep panels feed both diagnoses) and tech & attention (the digital substitution that ate the in-person hours is measured there).