默认身份：中国护照 + 长期在美。双榜并行：海归榜采用九维等权——生活成本、医疗资源、回国安顿便利、气候与自然环境、交通网络、吃喝日常、城市好玩度、子女教育、城市气质，总分为九维平均 × 10，其中生活成本采用北美参照系的梯次调整分；本地榜去掉回国安顿便利整维、生活成本采用原始分，总分为八维平均 × 10。两榜总分均显示一位小数，同分并列。维度得分 9–10 为强项、7–8 为中上、5–6 为中等、4 及以下为短板。回国安顿便利由国内长期安顿 29.41%、美国系统维护 26.47%、赴美便利 25.88%、专业服务 18.24% 四个子项构成。定制加权报告与美国版同一规则：“最看重”与“最不看重”的选择合计计数，合计 1 项免费生成报告，合计 2 项及以上为 $1.99。Default profile: Chinese passport + long-term U.S. resident. Two editions run side by side. The returnee edition equally weights nine dimensions — cost of living, healthcare, settling-back ease, climate and environment, transport, daily food and drink, city fun, children’s education and city temperament — with the total equal to the nine-dimension average × 10 and cost of living on the tiered North-America-reference scale. The local edition drops settling-back ease entirely and uses raw cost-of-living scores; its total is the eight-dimension average × 10. Totals on both editions are shown to one decimal place, with equal displayed scores tied. Dimension scores of 9–10 are strengths, 7–8 above average, 5–6 middling, and 4 or below weaknesses. Settling-back ease combines four sub-items: long-term settling in China 29.41%, U.S.-system upkeep 26.47%, U.S.-bound travel ease 25.88% and professional services 18.24%. The custom-weighted report follows the same rule as the U.S. edition: most- and least-important choices count together — 1 total choice generates a free report, 2 or more cost $1.99.
2026-10-18 最新更新：① 回国安顿便利删除「购房成本与难度」子项——它与生活成本中的住房成本重复计算；其余四子项权重按比例重分（国内长期安顿 29.41%／美国系统维护 26.47%／赴美便利 25.88%／专业服务 18.24%），北京回国安顿升至 9.34、上海升至 9.38、广州降至 8.44。② 海归榜生活成本按北美参照系做梯次加分——北京、上海 +3，深圳 +2，杭州 +1.5，南京、广州 +1，其余 34 城 +0.5，单项封顶 10 分。海归读者的价格锚点是北美：国内一线城市三居室月租仍不到湾区／纽约同档住房的一半，按中国 40 城内部排名得出的 1–2 分低估了这一体感。调整保留一线城市内部的梯度（广州 5.85 ＞ 南京 5.74 ＞ 杭州 5.43 ＞ 深圳 5.23 ＞ 北京 4.95 ＞ 上海 4.50），名次基本不变。③ 排名总分统一显示一位小数。海归榜 Top 5：广州 76.7 / 北京 75.4 / 上海 75.1 / 成都 74.8 / 杭州 74.5；本地榜 Top 5：广州 74.5 / 成都 73.7 / 杭州 72.3 / 武汉 71.1 / 西安 69.9。Latest update on 2026-10-18: (1) Settling-back ease drops its home-purchase cost-and-difficulty sub-item, which double-counted the housing costs already captured in cost of living; the remaining four sub-items are reweighted proportionally (long-term settling 29.41% / U.S.-system upkeep 26.47% / U.S.-bound travel 25.88% / professional services 18.24%) — Beijing’s settling-back score rises to 9.34 and Shanghai’s to 9.38, while Guangzhou’s eases to 8.44. (2) The returnee edition’s cost-of-living scores gain a tiered North-America-reference adjustment — Beijing and Shanghai +3, Shenzhen +2, Hangzhou +1.5, Nanjing and Guangzhou +1, and the other 34 cities +0.5, capped at 10. Returnee readers anchor prices to North America: a three-bedroom in China’s priciest first-tier cities still rents for less than half of a comparable Bay Area / New York home, so the 1–2 points those cities earned in a China-only comparison understated that reality. The adjustment preserves the first-tier gradient (Guangzhou 5.85 > Nanjing 5.74 > Hangzhou 5.43 > Shenzhen 5.23 > Beijing 4.95 > Shanghai 4.50) and leaves the ranking essentially unchanged. (3) Ranking totals are now shown to one decimal place. Returnee Top 5: Guangzhou 76.7 / Beijing 75.4 / Shanghai 75.1 / Chengdu 74.8 / Hangzhou 74.5. Local Top 5: Guangzhou 74.5 / Chengdu 73.7 / Hangzhou 72.3 / Wuhan 71.1 / Xi'an 69.9.
2026-10-18 更新：本版汇总当日六项调整，40 城总分与排名全量重算。① 吃喝日常重构落盘：咖啡奶茶 15% / 高端餐饮 23% / 街头饮食 17% / 生活便利度 25% / 生鲜食材 20%，删除「家乡味」子项，口碑修正与咖啡奶茶城市能级修正一并生效，烟台吃喝升至 4.77。② 城市好玩度拆为原有指标 70% + 城区休闲密度 30%（酒吧密度、Livehouse／小剧场、第三空间三项等权）：郑州好玩度 8.3→7.78，长沙休闲密度 8.33 居全国第二。③ 胶东沿海气候统一修正 ERA5 沿海偏差：烟台热不适 58.2→38 天/年、威海 36.2→30 天/年、青岛 41.8→35 天/年。④ 夜生活子项改用四档制（按小红书笔记密度定档：重庆 10 / 长沙 9.5 / 成都 9 / 广州 8.5；上海、北京 7.5；青岛 7；烟台 6；威海、大连、三亚 4），17 城好玩度随之重算。⑤ 城市气质新增体感修正层：最终气质分 = 数据分 + 站内笔记净情感 × 0.5，修正幅度上限 ±0.5——郑州 −0.18（7.99）、长沙 +0.20、广州 +0.10、杭州 −0.08、烟台 −0.15、玉溪 +0.05。⑥ 气候权重调整：体感舒适 50%→55%、自然灾害风险 18.75%→13.75%（空气质量 31.25% 不变）；灾害项对农业干旱等「区域致灾、城区无直接影响」的事件打折计算，昆明灾害分 6.73→7.5、气候总分 7.43→7.60。Update on 2026-10-18: this edition folds in six same-day adjustments and recalculates all 40 totals and ranks. (1) The daily food & drink rebuild is now live — coffee & tea 15% / high-end dining 23% / street food 17% / everyday convenience 25% / fresh ingredients 20%, with the former “taste of home” sub-item removed; reputation adjustments and the coffee & tea city-tier correction are included, and Yantai’s food score rises to 4.77. (2) City fun is now 70% existing indicators + 30% urban leisure density (bar density, livehouses / small theatres and third places, equally weighted): Zhengzhou’s fun score falls 8.3 → 7.78, while Changsha’s leisure density of 8.33 ranks second nationwide. (3) Jiaodong coastal climate scores were unified to correct an ERA5 coastal bias — Yantai heat-discomfort days 58.2 → 38 per year, Weihai 36.2 → 30, Qingdao 41.8 → 35. (4) The nightlife input now uses a four-tier scale based on Xiaohongshu note density (Chongqing 10 / Changsha 9.5 / Chengdu 9 / Guangzhou 8.5; Shanghai and Beijing 7.5; Qingdao 7; Yantai 6; Weihai, Dalian and Sanya 4), recalculating city fun for 17 cities. (5) City temperament gains a perception-calibration layer: final temperament = data score + on-platform net sentiment × 0.5, capped at ±0.5 — Zhengzhou −0.18 (7.99), Changsha +0.20, Guangzhou +0.10, Hangzhou −0.08, Yantai −0.15, Yuxi +0.05. (6) Climate weights were rebalanced: thermal comfort 50% → 55% and natural-hazard risk 18.75% → 13.75% (air quality unchanged at 31.25%); hazard events that strike a region without directly affecting the urban core, such as agricultural droughts, are now discounted — Kunming’s hazard score rises 6.73 → 7.5 and its climate total 7.43 → 7.60.
2026-10-17 更新：吃喝日常子项重构——咖啡奶茶 15% + 高端餐饮密度 23% + 街头饮食生态 17% + 生活便利度 25% + 生鲜食材 20%，原「家乡味」子项删除。生活便利度由快递、便利店、商业活力、外卖、高品质商超五子项等权构成；生鲜食材细化水产四档、昆明地标食材上调、电商补充本地平台；街头饮食采用高德扫街榜四档制，并对东北城市增加烧烤／地方小吃密度修正（沈阳街头分 2.5→3.75）。长沙必吃榜按 46 家实测计、水产上调至 7.5 分，杭州必吃榜按 61 家实测计。40 城总分与排名已全量重算。Update on 2026-10-17: the daily food & drink dimension was rebuilt — coffee & tea 15% + high-end dining density 23% + street-food ecology 17% + everyday convenience 25% + fresh ingredients 20%; the former “taste of home” sub-item was removed. Everyday convenience now averages five equal inputs (courier volume, convenience stores, commercial vitality, food delivery and premium supermarkets). Fresh ingredients refine seafood into four tiers, lift Kunming’s landmark-ingredient score and add local platforms to e-commerce coverage. Street food uses Amap’s four-tier “sweep-the-streets” list, with a barbecue / local-snack density adjustment for northeastern cities (Shenyang’s street score 2.5 → 3.75). Changsha’s must-eat count uses the observed 46 restaurants and its seafood score rises to 7.5; Hangzhou uses the observed 61. All 40 totals and ranks were recalculated.
2026-10-16 更新：自然资源并入城市好玩度，采用加分制——人文娱乐保留 100%，自然资源高于 5 分的部分按（自然分−5）÷2 加分，最高 +1，城市好玩度上限 10。这样保留人文与自然各自的长板，不再单列自然资源维度。Update on 2026-10-16: natural assets now contribute a bonus to city fun. The culture-and-entertainment score remains intact; natural assets above 5 add (nature − 5) ÷ 2, capped at +1, and city fun is capped at 10. Natural assets are no longer a standalone dimension.
2026-10-16 五项口径更新：① 市民素质改用官方文明建设认证代理指标并完成 40 城全覆盖（34 城 8.0 分、2 城 7.0 分、4 城 5.0 分）；② 道路规划（交通健康度）完成 40 城全覆盖——高德 27 城实测、百度地图百城指数补齐其余 13 城（其中 6 城为估算指数），两套指数经锚点映射统一归一化，合并与分开计算结果一致；③ 空气质量改用 PM2.5 主导权重（PM2.5 65% / 臭氧 22.5% / NO₂·SO₂ 12.5%），其余污染物暂无统一数据时按 PM2.5 分量归一化，三亚空气质量升至 9.4 分；④ 体感舒适改用含湿度的热不适天数（热指数 HI≥95°F，ERA5 2021–2025）替代只数 ≥32°C 天数——三亚热不适天数由旧口径 10.4 天/年修正为 182.8 天/年，湿热城市体感分普遍下调；⑤ 北京子女教育参数调整：卷度惩罚 0.25→0.22、基础教育原始分 6.8→7.2，北京子女教育升至 8.0 分，与上海（8.2 分）的差距收窄至 0.2 分。以上变更后，40 城总分与排名已全量重算，全部 40 城进入正式排名。Five methodology updates on 2026-10-16: (1) citizen quality is now proxied by official civility-certification results, covering all 40 cities (34 cities at 8.0, 2 at 7.0, 4 at 5.0); (2) road planning (traffic health) now covers all 40 cities — 27 observed from Amap and 13 completed from Baidu Maps’ 100-city index (6 of them estimated) — with both indices anchor-mapped onto one normalized scale, verified to give identical results whether merged or computed separately; (3) air quality now uses PM2.5-dominant weights (PM2.5 65% / ozone 22.5% / NO₂·SO₂ 12.5%), normalized on the PM2.5 component where the other pollutants lack unified data, lifting Sanya’s air quality to 9.4; (4) thermal comfort now counts heat-discomfort days on a humidity-inclusive heat index (HI ≥ 95°F, ERA5 2021–2025) instead of days ≥ 32°C alone — Sanya’s heat-discomfort days are revised from 10.4 to 182.8 per year, lowering comfort scores across humid southern cities; (5) Beijing’s education parameters were adjusted (competition penalty 0.25 → 0.22; base-education raw score 6.8 → 7.2), lifting Beijing’s children’s-education score to 8.0 and narrowing its gap with Shanghai (8.2) to 0.2. All 40 totals and ranks were recalculated, and all 40 cities now enter the official ranking.
城市气质替换原社交友好度，由官方文明建设认证（市民素质代理指标）、服务意识、道路规划（交通健康度）三项各占三分之一；公共场所禁烟作为额外加分。道路规划采用高德交通健康指数（27 城）与百度地图百城通勤拥堵指数（13 城）两套来源，经锚点映射统一归一化到 0–10（最堵 0 分、最不堵 10 分），反映交通运行状况，不直接衡量道路设计或步行环境。外地人融入度与方言友好度保留为定性备注，不直接计分。City temperament replaces social friendliness. Official civility certification (the citizen-quality proxy), service awareness and road planning (traffic health) each contribute one third, with an additional smoke-free-public-spaces bonus. Road planning combines Amap’s traffic-health index (27 cities) with Baidu Maps’ 100-city commuting congestion index (13 cities), anchor-mapped onto a single 0–10 scale (most congested 0, least congested 10); it reflects operating conditions, not road design or walkability. Integration and dialect friendliness remain qualitative notes rather than scored inputs.
城市气质数据覆盖：官方文明建设认证与道路规划两个子项已覆盖全部 40 城；服务意识有 22 城直接数据，其余 18 城暂无公开数据，按已有子项归一化计分。全部 40 城均进入正式排名。贵阳、南宁、襄阳 3 城的自然灾害风险数据已完成补充核验（2026-10-16），“待核验”标记解除。City-temperament coverage: the civility-certification and road-planning inputs now cover all 40 cities; service awareness has direct data for 22 cities, while the remaining 18 are normalized across available inputs. All 40 cities enter the official ranking. Supplementary verification of natural-hazard-risk data for Guiyang, Nanning and Xiangyang was completed on 2026-10-16 and their “Pending review” flags were removed.
2026-10-16 医疗地理口径修订：“每百万人三甲医院数”的人口分母由全市常住人口改为住建部年鉴口径的城区人口，减少大行政区城市被系统性低估的问题；29 城采用实测城区人口，11 城暂用估算值并待年鉴核实。此前发现的 10 倍单位错误已修正，档位继续采用：≥4.0 家为 10 分、3.0–4.0 为 8 分、2.0–3.0 为 6 分、1.5–2.0 为 4 分、1.0–1.5 为 2 分、<1.0 为 0 分。医疗资源与九维总分已全量重算。Healthcare geographic-scope revision on 2026-10-16: the population denominator for Grade A tertiary hospitals per million residents changed from total municipal population to the urban-district population used by China’s Ministry of Housing and Urban-Rural Development yearbook, reducing the systematic understatement of large administrative cities. Twenty-nine cities use observed urban-district populations; 11 use provisional estimates pending yearbook verification. The previously identified 10× unit error remains corrected, with scoring bands unchanged: ≥4.0 scores 10; 3.0–4.0 scores 8; 2.0–3.0 scores 6; 1.5–2.0 scores 4; 1.0–1.5 scores 2; and <1.0 scores 0. Healthcare and all nine-dimension totals were recalculated.
气候与自然环境由空气质量 31.25% / 自然灾害风险 13.75% / 体感舒适 55% 构成（2026-10-18 起体感权重由 50% 上调、灾害权重由 18.75% 下调）。自然灾害风险由气象灾害风险与地质灾害风险按 2:1 加权合成，纳入地震危险性（GB 18306-2015 区划）与近 10 年实际灾害事件；原“极端天气”子项自本版起由该子项取代。体感舒适月＝基础舒适月份（月均温 ≥10°C 且月均高温 <90°F）−暖季阴雨−闷热。Climate and environment combines air quality 31.25%, natural-hazard risk 13.75% and thermal comfort 55% (from 2026-10-18, comfort is upweighted from 50% and hazard risk downweighted from 18.75%). Natural-hazard risk weights meteorological and geological hazard risk 2:1, incorporating seismic hazard (GB 18306-2015 zoning) and actual disaster events over the past decade; it replaces the former “extreme weather” sub-item from this edition onward.
交通口径：国际直飞航点数不含北美，港澳台按 0.5 个航点计、其余国际航点按 1 个计；市内轨道按人口规模归一化比较。Transport scoring excludes North America from international nonstop destinations; Hong Kong, Macao and Taiwan count as 0.5 destination each, while other international destinations count as 1. Urban rail is normalized by population.
无公开数据的子项暂不计分，维内按其余子项归一化并标注“数据待补充”。沈阳与大理三居室租金已补齐；天津仍以读者实测作参考，义乌出租车起步价暂不计分。Numbeo 仅作第三方交叉参考，不进入本榜公式。Unpublished sub-items are left unscored and dimensions are normalized across available inputs. Three-bedroom rents are now filled for Shenyang and Dali; Tianjin still uses a reader-reported reference, while Yiwu's taxi base fare remains unscored. Numbeo is shown only as a third-party cross-reference and is not scored.
