Bus–Metro–Road: Data from Indian Cities
- Neural City Team
- Sep 12
- 11 min read
Updated: Oct 7
India’s big-city congestion is usually debated in silos—“we need more metros,” “we need more buses,” “we need better footpaths.”
For this story our brief was to put these pieces on one page.
Assemble a comparable view of bus fleets, metro kilometres, and road supply across major Indian cities and global high-density references, Tokyo and Beijing.
The second question sitting behind the numbers was walkability: there’s a lot of chatter around it, but walkability only works when public transport is worth walking to—and strong public transport works better when the last 500 metres are walkable. The two rise and fall together over the long term.
The first pass at the data threw a curveball.
Bengaluru looked better than peers on several counts—plenty of road length, the strongest buses per capita in our Indian set—yet it is the byword for congestion.
The numbers alone weren’t completing the story. A quick look at the city map clarified why: road length is not the same as a road network. Bengaluru’s totals hide narrow average widths and discontinuities that choke junctions and corridors; what you gain in kilometres, you lose in throughput.
City | Population | City Area | Population density | Buses per 100k people | Public city buses (units) | Operational metro length (km) | Road network length (km) | Road Density | Avg time to drive 10 km at peak | Avg Speed (km/h) | Most-used public transport | Daily Ridership (Mn) | GDP per capita (Rs) | Privately owned Cars (Mn) | Cars per 100,000 People | Cars per sq km | Estimated avg road width (m) |
Delhi (NCT) | 3,46,66,000 | 1,483 | 23,376 | 22 | 7683 | 395 | 18460.4 | 12 | 28 | 21 | Metro | 5.6 | 501000 | 2070000 | 5971 | 1396 | 24 |
Mumbai | 2,20,89,000 | 603 | 36,632 | 12 | 2758 | 69 | 2000 | 3 | 37 | 16 | Suburban Rail | 6.2 | 657000 | 1400000 | 6338 | 2322 | 18 |
Bengaluru | 1,43,95,000 | 709 | 20,303 | 49 | 7000 | 80 | 12878.8 | 18 | 40 | 16 | Buses | 4.5 | 893000 | 2310000 | 16047 | 3258 | 12 |
Chennai | 1,23,36,000 | 426 | 28,958 | 27 | 3375 | 54 | 5657.33 | 13 | 30 | 15 | Buses | 3.2 | 20 | ||||
Hyderabad | 1,13,38,000 | 650 | 17,443 | 28 | 3151 | 69 | 9013 | 14 | 27 | 22 | Buses | 2.5 | 20 | ||||
Pune | 75,26,000 | 527 | 14,281 | 25 | 1916 | 33 | 2044 | 4 | 40 | 15 | Buses | 1.2 | 12 | ||||
Gurugram | ~20,00,000 | 232 | 5,526 | 12 | 150 | 13 | 4596 | 20 | 35 | 15 | Buses | 0.09 | 30 | ||||
Beijing | 2,18,32,000 | 1,500.00 | 14,555 | 105 | 23000 | 879 | 9027.17 | 6 | 27 | 22 | Metro | 9.7 | 10 | ||||
Tokyo (23 wards & metro) | 1,42,54,000 | 627.51 | 22,715 | 10 | 1467 | 304 | 12,012 | 19 | 25 | 24 | Metro | 9.4 | 10 | ||||
Jakarta | 1,06,85,000 | 662.33 | 16,132 | 40 | 4300 | 16 | 6936 | 10 | 40 | 15 | Buses | 1.01 | 12 |
Baseline comparisons
Roads Density (km of road / sq km of city) — highest
Gurugram 20 km
Bengaluru 18 km
Hyderabad 14 km
Roads (km) — max supply
Delhi (NCT) 18,460 km
Bengaluru 12,878 km
Hyderabad 9013 km
Metro (operational km) — most supply
Delhi (NCT) 395 km
Bengaluru 80 km
Hyderabad 69 km
Metro (operational km) — least supply
Gurugram 12.85 km
Pune 33.1 km
Chennai 54 km
Buses per 100k people — highest
Bengaluru 49 (highest in India in our set)
Hyderabad 28
Chennai 27
Peak 10-km time (TomTom) — slowest
Pune 41 min
Gurugram 40 min
Bengaluru 38 min(Delhi is best among Indian cities at ~28 min; Tokyo 25; Beijing 27.)
Level-2 insights what it implies
1) Rails move the masses; buses are feeders
In high-density cities, grade-separated rail (metro/suburban) carries far more people than buses.
India check: Delhi is the only city here with metro scale approaching comprehensive coverage (~11.4 km per million people). Others sit at ~4–6 km/million (Mumbai, Chennai, Pune, Bengaluru, Hyderabad), which forces more demand back onto streets.
Global contrast: Tokyo (~21.3 km/million) and Beijing (~40.3 km/million) show how much rail you need for dense cores; buses then work best as high-frequency feeders to rail, not as trunk movers.
2) Bengaluru’s paradox: lots of roads and buses, still jammed
Road supply looks OK on paper:
~1,816 road-km per 100 km² (second among Indian cities here, after Gurugram).
Bus supply is strong: ~49 buses per 100k people (top among Indian cities in this set), yet peak time is ~38 min for 10 km.
But geometry limits throughput:
Network form matters: Satellite/OSM view shows discontinuous ring/radial structure—long orbital gaps and few continuous cross-city links. Road length exists, but network connectivity is weak, so flows collapse into a handful of arterials.
~12 m average street width concentrates conflicts at intersections; curb friction is high.
Rail is under-scaled for its load:
~5.5 metro-km per million—better than some, but far behind Delhi and drastically below Tokyo/Beijing.
What that means: Without continuous orbitals/radials, bus lanes + signal priority, and faster metro build-out, added buses just queue in the same bottlenecks.
City Implications
Mumbai — Rail must carry the city
Road scarcity: Just 9.1 road-km per 100k people (lowest in the set) and ~332 road-km per 100 km².
Street geometry: average width ~18 m; density a crushing ~36.6k/km².
Rail dependence: Metro km per million people is ~3.1, but suburban rail is the real backbone (very large daily ridership), dwarfing bus capacity.
What this implies:
Low total road length and very high density mean streets can’t carry the city; the backbone has to be rail. Metro coverage is modest while the suburban network does the heavy lifting. Treat metro + suburban scale-up as the primary capacity lever (Tokyo is the relevant peer: narrow roads, heavy rail).
Delhi (NCT) — Balanced supply, better outcomes
Road abundance: 17,882 km total; ~1,245 road-km per 100 km²; width ~24 m.
Rail supply: 395 km metro; ~11.4 km per million people.
Result: Among India’s best peak 10-km time (~28 min) despite high density.
Implication:
Among the best Indian metrics: large road network, relatively wider streets, and the biggest metro. Still, when set against Beijing/Tokyo, Delhi’s rail intensity per capita and per area can go further. Metro expansion should continue to close the gap with top Asian systems.
Bengaluru — Plenty of road length, but narrow streets and rail gap
Road density: a striking ~1,816 road-km per 100 km² (high), but avg width ~12 m limits throughput.
Bus strength: 49 buses/100k (top in India here).
Rail supply: ~80 km metro → ~5.5 km per million people (mid-table).
Outcome: Peak 38 min/10 km despite strong buses.
Implication:
Plenty of road kilometres but narrow average widths; highest buses per capita in the Indian set, yet peak travel time remains high. Metro length is mid-table. The data points to a rail shortfall on trunk corridors; buses are carrying a lot on constrained streets.
Hyderabad — Solid balance, quietly efficient
Road density: ~1,503 road-km per 100 km², ~20 m width.
Rail supply: ~69 km metro → ~6.1 km per million
Outcome: Best Indian peak time (~27 min) in your set.
Implication:
Good road supply and moderate density align with the best peak travel time among Indian cities here; metro length is solid but not top. Maintain the edge by continuing metro build-out alongside existing road capacity.
Chennai — High density, moderate rail, heavy suburban support
Density: ~28.9k/km²; road density ~1,328 km/100 km²; width ~20 m.
Rail: ~54 km metro (~4.4 km/million), plus suburban/MRTS (sizeable riders in your table).
Outcome: 39 min/10 km
Implication:
Strong citywide road footprint and high density; metro length is moderate and peak time is slower than Delhi. Keep metro expansion on track and ensure bus–metro connectivity remains tight to convert that road footprint into throughput.
Pune — Tight streets, low rapid transit, worst peak time
Road geometry: ~12 m avg width; ~394 road-km per 100 km² (low).
Rail: ~33 km metro (~4.4 km/million)—small for the load.
Outcome: 41 min/10 km (worst).
Implication:
Metrics are not in its favour: thin streets, short metro, lower bus supply, and the worst peak 10-km time in the set. Public-transport capacity is undersized across modes.
Gurugram — Widest roads, thinnest bus service
Street geometry: ~30 m avg width; ~1,981 road-km per 100 km² (very high).
Transit supply: 12.85 km metro; only ~12 buses/100k.
Outcome: 40 min/10 km peak despite wide roads.
Implication:
Widest streets and high road-km density, but lowest bus supply and minimal metro length; peak time is still among the worst. Width without transit supply equals congestion—bus inventory and corridor coverage are the missing pieces.
Beijing & Tokyo (what high-density readiness looks like)
Beijing: Narrow streets (~10 m) and medium road density, but huge metro per capita (~40 km/million) and very high buses/100k (105) → 27 min/10 km.
Tokyo (23 wards): Very narrow (~10 m) yet rail-dominant (metro + JR/private) → 25 min/10 km.
Implication: In very dense cities, rail scale + bus precision + walking quality beat width. That’s the North Star for Mumbai/Bengaluru
The above comparison leaves us with two practical takeaways. First, if you want a single “readiness” lens, it’s the triad: road supply (length and typical width) + mass-rapid rail supply (metro/suburban) + surface transit supply (buses).
The mix—not any one number—explains why Delhi outperforms on peak travel time, why Mumbai must lean on rail like Tokyo, and why Gurugram’s wide roads still stall without buses and metro coverage.
Second, totals don’t tell geometry. A city can have lots of road kilometres and still move slowly if the network is narrow, broken, or forces long detours; a simple map panel alongside the table prevents false comfort.
Data Sources
Bus
Mumbai — BEST fleet 2,758 (May 8, 2025, Hindustan Times) Hindustan Times; metro length compiled from system pages (cross-checked with The Metro Rail Guy, May 2025). Motor India Online
Delhi (NCT) — Total city buses ~7,683 (DTC/Delhi bus page; aligns with recent reports on DTC+Cluster totals); metro ≈390–395 km network. Wikipedia+1Bharat Metro Locator
Bengaluru — BMTC “7,000+ buses” (Sep 2025, Times of India); Namma Metro operational ≈79.65 km (Sep 2025). The Times of IndiaHousing
Hyderabad — TSRTC GHMC fleet 3,151 (Aug 2025, TOI); Hyderabad Metro length ≈69.2 km (official L&T Metro). The Times of IndiaLT Metro
Chennai — MTC fleet 3,375 (Tamil Nadu Transport Dept “Performance” page, Aug 2025); CMRL operational ≈54.1 km (CMRL/Wikipedia). MTC BusWikipedia
Pune — PMPML fleet 1,916 (Feb 2025, Indian Express); Pune Metro operational ≈33.1 km (Bharat Metro Locator, Jul 2025). The Indian ExpressBharat Metro Locator
Gurugram (GMCBL/Gurugaman) — ~150 city buses currently in operation (2025 coverage; TOI/HT). Rapid Metro length ≈12.85 km. The Times of India+1Wikipedia
Beijing — “>23,000 buses” in fleet; ~20,000 in daily operation (Beijing Gov English portal, Oct 2024); Subway ≈879 km (2025). Beijing Government PortalWikipedia
Tokyo — Toei Bus fleet 1,467 (major public operator; private companies not included); Tokyo Subway (Metro+Toei) ≈304.1 km. Wikipedia+1
Jakarta — TransJakarta fleet ~4,300; MRT operational 15.7 km. Wikipedia+1
Population basis used:
For most Indian cities: UN/World-Population-Review or Macrotrends urban-agglomeration estimates (2025) to keep a uniform method across metros (e.g., Delhi, Bengaluru, Hyderabad, Chennai, Pune, Mumbai). World Population Review+2World Population Review+2Macrotrends+2Macrotrends+2
Beijing: City proper (Beijing Municipal Government 2024). Beijing Government Portal
Tokyo: Tokyo Metropolis (city proper, May 2025). Note that bus count shown is Toei Bus only; many private operators (Keio, Odakyu, Tokyu, etc.) are not included, so Tokyo’s total buses are higher than shown. Wikipedia+1
Jakarta: City proper (BPS mid-2024). Wikipedia
Gurugram: City proper estimate for 2025 as census was postponed. (Census2011.co.in projection.) Census 2011 India
Important caveats:
Scope = intra-city public buses. For India I used the main city operators (BEST, DTC/Cluster, BMTC, TSRTC (GHMC), MTC, PMPML, GMCBL). This excludes state intercity buses and private minibuses unless they are part of the city operator’s fleet.
Tokyo’s bus figure reflects Toei Bus only (public operator). Tokyo has multiple large private bus companies; adding them would materially increase the total, but comprehensive, current counts are not consolidated publicly.
Populations differ by definition (city proper vs. UA). I chose a consistent approach per city as noted in the Population basis column so your per-capita calculations are reproducible.
Metro length is operational route length (opened to the public), excluding suburban/commuter rail. For Gurugram I included Rapid Metro; Delhi’s figure is for Delhi Metro (NCR). Beijing/Tokyo are their subway/metro only.
Road Length (km):
Mumbai (BMC): 2,050 (BMC Statistical Year Book 2024–25). MyBMC
Delhi (NCT): 17,882 (Economic Survey of Delhi 2021–22; total across agencies). Delhi Planning Department
Bengaluru (BBMP): 12,878.84 (BBMP figure reported Feb 23, 2025). Bangalore Mirror
Hyderabad (GHMC): 9,013 (GHMC jurisdiction, Mar 2025).
Chennai (GCC): 5,657.33 (5,270.33 km interior + 387 km bus-route roads; GCC site). Chennai Corporation
Pune (PMC): 2,044 (PMC figure reported July 17, 2025). Instagram
Gurugram (MCG): 4,596 (MCG jurisdiction; multiple reports 2024–25). The Times of Indiagurgaonfirst.org
Beijing (urban roads): 6,256 (end-2023 urban road length; official communiquè). beijing.gov.cntjj.beijing.gov.cn
Tokyo (Metropolis): 24,741 (as of Apr 1, 2022; entire Tokyo Metropolis incl. 23 wards + western cities). kensetsu.metro.tokyo.lg.jp
Jakarta (DKI): 6,492 (2022 total; BPS-cited in ATO “State of Play”). Asian Transport Observatory
Notes on scope & comparability (important):
India figures are for municipal corporation or NCT/MCG limits (not whole metro areas).
Beijing: I used the urban road length at end-2023; the same communiqué also lists total highways across the municipality (22,433.2 km), which is broader. beijing.gov.cntjj.beijing.gov.cn
Tokyo: official total is for the entire Tokyo Metropolis (includes the 23 wards plus outlying cities). If you prefer 23-wards-only, TMG doesn’t publish a single headline km figure for just the wards; we can estimate from ward-level data, but the metropolis total above is the authoritative published number. kensetsu.metro.tokyo.lg.jp
Jakarta: latest clearly cited total is 2022 (6,492 km). BPS has a 2023 municipal breakdown page, but it’s rendered dynamically; the ATO doc is explicit and cites BPS as the source. jakarta.bps.go.idAsian Transport Observatory
Metro
City | Public city buses (units) | Population (millions) | Population basis | Buses per 100k people | Operational metro length (km) | Key sources |
Mumbai | 2758 | 22.089 | Urban agglomeration (UN/WPR 2025) | 12.4858527 | 68.93 | Hindustan Times (May 8, 2025); The Metro Rail Guy (May 9, 2025); Wikipedia Mumbai Metro (May 2025) |
Delhi (NCT) | 7683 | 34.666 | Metro area (Macrotrends 2025) | 22.1629262 | 395 | Wikipedia DTC; Macrotrends (Delhi metro area 2025); Wikipedia Delhi Metro (Aug 2025) |
Bengaluru | 7000 | 14.395 | Metro area (Macrotrends 2025) | 48.6279958 | 79.65 | Times of India (Sep 2025) for BMTC ~7,000; Housing.com news (Sep 2025) for 79.65 km |
Hyderabad | 3151 | 11.338 | Urban agglomeration (WPR 2025) | 27.7914976 | 69.2 | TOI (Aug 2025) fleet 3,151; L&T Metro/Hyderabad Metro official (Jan 2025) 69.2 km |
Chennai | 3375 | 12.336 | Urban agglomeration (WPR 2025) | 27.3589494 | 54.1 | TN MTC Performance (Aug 2025) 3,375 buses; Wikipedia/CMRL (2021–2025) 54.1 km |
Pune | 1916 | 7.526 | Metro area (Macrotrends 2025) | 25.4584108 | 33.1 | Indian Express (Feb 2025) 1,916 buses; Bharat Metro Locator (Jul 2025) 33.1 km |
Gurugram | 150 | 1.282 | City proper (Census2011/est. 2025) | 11.700468 | 12.85 | TOI/HT (2024–2025) ~150 buses in city; Wikipedia Rapid Metro 12.85 km; census2011.co.in estimate 2025 |
Beijing | 23000 | 21.832 | City proper (Beijing Govt 2024) | 105.349945 | 879 | Beijing Govt (Oct 2024) >23k fleet, 20k in operation; Wikipedia Beijing Subway (Sep 2025) 879 km |
Tokyo (23 wards & metro) | 1467 | 14.254 | City proper (Tokyo Metropolis 2025) | 10.2918479 | 304.1 | Toei Bus (Wikipedia): 1,467 buses; Tokyo Metro + Toei Subway (Wikipedia) 195.1+109.0=304.1 km; Tokyo Metropolis pop (May 2025) |
Jakarta | 4300 | 10.685 | City proper (BPS mid-2024) | 40.2433318 | 15.7 | Wikipedia TransJakarta (Aug/Sep 2025) ~4,300 buses; BPS Jakarta mid-2024 pop; Wikipedia Jakarta MRT 15.7 km |
Daily Ridership
Area | Most-used public transport | Daily ridership (approx.) |
Mumbai | Suburban rail (“local trains”) | ~6.2 million/day in FY2022–23; approaching pre-COVID levels by late-2024 reporting. (Wikipedia) |
Delhi (NCT) | Delhi Metro | ~5.6 million/day in 2023 (annual 2.03 billion/yr); single-day record 8.19 million on Aug 8, 2025. (Wikipedia) |
Bengaluru | BMTC city buses | ~4.4–4.6 million/day after the Shakti scheme boost (mid-2025). (Deccan Herald) |
Hyderabad | TSRTC city buses (Hyderabad–Secunderabad) | ~2.5 million/day in 2025 (city region); state-wide TSRTC ~5.4–6.0 million/day. (The Times of India) |
Chennai | MTC city buses | ~3.1–3.4 million/day in 2024–25 (official performance page). (MTC Bus) |
Pune | PMPML city buses | ~1.2 million/day in 2023–24 (with recent fluctuations in 2025). (The Times of India) |
Gurugram | Gurugaman city buses | ~90,000/day (2023–24 reports); Rapid Metro is much smaller (~30k–48k/day). (The Times of India) |
Beijing | Beijing Subway | ~7.8–9.7 million/day (weekly average range, late-Aug 2025). (MacroMicro) |
Tokyo (23 wards & metro) | Subway (Tokyo Metro + Toei) | ~9.4 million/day in FY2023 (Tokyo Metro 6.52m + Toei 2.85m). (Wikipedia) |
Jakarta | TransJakarta BRT | ~1.01 million/day average in 2024 (peaks ~1.3 million). (Wikipedia) |
Road Width and Taxis
City | Road Width Major/Avg (m) | Source/Assumption (Width) | Taxis (approx) | Source/Assumption (Taxis) | Key Caveats |
Mumbai | 18–45 (major) | Indian urban arterials/local standards, Mumbai DP 2034, ITDPduac | 55,000 | Regulator, aggregator, & legacy taxi permits, 2024–25 news reports | Includes old/KAALI-PEELI and radio cabs |
Delhi | 24–45 (major) | DUAC/IRC guidelines, arterial & ring road referencesduac | 100,000 | Delhi Transport Dept., Taxi Board, Ola/Uber registered, estimates | Radio cabs, regular, includes some NCR overlap |
Bangalore | 12–30 (common) | Municipal planning docs, arterial/collector categories; major >24mduac+1 | 80,000 | City traffic police, ride-hailing operators’ 2025 estimates | Includes ride hailers (app-based) |
Hyderabad | 30–60 (major) | GHMC/urban arterial guidelines, project road docstimesofindia.indiatimes | 50,000 | Telangana permit data, local taxi association, Ola/Uber estimates | Most operate mainly in urban Hyderabad |
Chennai | 11–36 (arterial/major) | Chennai Corporation, DP, local & arterial street mappingtimesofindia.indiatimes+1 | 30,000 | Red Taxi, city records, Ola/Uber + taxi associations | Includes large radio taxi fleet |
Pune | 12–18 (arterial, city) | PMC, local road projects, ITDP street guidelinesitdp | 20,000 | Taxi app provider data, city news, PMPML, with app cabs | Huge app-cab share, few traditional city taxis |
Gurgaon | 12–75 (sector/arterial) | Haryana PWD, GMDA documents (new sectors: up to 75m)timesofindia.indiatimes+1 | 12,000 | Transport, taxi operator data, news, NCR taxi registry | Cross-border, figures overlap with Delhi often |
Beijing | 7.3–10.4 (arterial avg.) | Municipal planning, research (urban Beijing, trunk ≈10m) acsa-arch+2 | 67,000–70,000 | Beijing Govt., Taxi Regulator (2024), China Daily, expat guides | Taxi licensure centralized, accurate recent data |
Tokyo | 10–15 (major arterials) | Urban planning/NPA, NACTO Urban Street Guide, Tokyo transit white paperstatista | 35,000 | Tokyo Taxi Association, annual metro data (2023–2024) | Large share are private hire, licensure strict |
Jakarta | 7–15 (road/lane avg) | JICA/Jakarta planning docs, arterial/major masterplanopenjicareport.jica+1 | 24,000–27,000 | Jakarta Transport Agency, news sources, local taxi operators | Excludes much of commuter metro taxi traffic |
Comments