Lit for Cars, Dark for People: Why Indian Streets Empty After 9 pm

Many Indian streets that are crowded at noon fall quiet soon after the main shops pull down their shutters. The usual response is more and brighter poles. This article examines why a street that works by day can fail at night, drawing on research on street lighting and crime, women's safety audits in Indian cities and laboratory work on how pedestrians read one another after dark. It argues that the night-time street is a question of urban form as much as illumination, and proposes a five-part field diagnostic, NIGHT-5, that ward engineers, designers and community audits can apply on foot.

Why the street after dark matters now

For many urban Indians the working day does not end at sunset. Shift workers, students, domestic workers, nurses and shop staff walk home and wait for buses well after 9 pm. Yet a familiar pattern repeats across cities: a market street that is dense with people, signs and goods in the afternoon becomes a corridor of shutters by night, with light concentrated on the carriageway and darkness along the footpath.

The usual response is to count poles and lamps, which treats the night as a lighting problem alone. This article argues that the night-time public realm depends on four things that urban design controls: which frontages stay open and lit, whether the walk is continuous, whether there are people and late uses within sight, and whether light reaches faces rather than only the road surface. The sections below review the evidence, read two drawn comparisons, set out the five tests, and close with limits and implications for practice.

What the research says

Jacobs (1961) argued that a street feels safe when there is a clear line between public and private space, when buildings face the street so that people inside can watch it, and when the footpath is used fairly continuously. Her "eyes on the street" were not cameras or guards but shopkeepers, residents and passers-by with a reason to be there. Gehl (2010) makes a related case for the city at eye level, arguing that open, lively ground floors and the presence of people help streets feel safe, including in the evening.

Lighting research supports the idea that light helps, but it complicates the reason why. In a systematic review and meta-analysis of 13 evaluations from high-income countries, Welsh and Farrington (2008) found that improved street lighting was associated with significant reductions in crime. They also reported that night-time crimes did not fall more than daytime crimes. The authors read this as support for a community explanation, in which visible investment strengthens pride and informal social control, rather than a simple deterrence effect of brightness.

Indian evidence has come largely from women's safety audits. Viswanath and Basu (2015) describe SafetiPin, a mobile application that turns the participatory safety audit into crowdsourced data, and report its use in seven Indian cities to identify the conditions linked with feeling unsafe. The audit rates nine parameters: lighting, openness, visibility, people, security, walk path, public transport, gender usage and feeling (Safetipin, 2019). Two definitions are telling. Lighting counts light from houses, shops and street vendors as well as street lamps, and visibility is explicitly framed as "eyes on the street" (Safetipin, 2019, pp. 9, 13).

The Delhi assessment sponsored by the city government's Department of Women and Child Development reports that lighting was rated 1.9 out of 3 and that 2,768 audit points, about 14%, were dark spots; this was down from 7,438 dark spots recorded in 2016, after agencies repaired and added lights (Safetipin, 2019, pp. 9–10). Visibility, however, was the weakest physical parameter at 0.8 out of 3, with 51% of points rated poor. The report attributes this mainly to high boundary walls that cut the visual link between pedestrians and buildings, and notes that wards with roadside commerce and vendors scored better (Safetipin, 2019, p. 13). Read together, these figures suggest that a city can fix lamps faster than it can fix blind edges.

Finally, laboratory work clarifies what pedestrian light is for. Fotios et al. (2015) asked participants to judge facial expressions under different lamp types, light levels and simulated distances. Accurate judgements needed a face luminance of roughly 0.1–1.0 cd/m² at 4 m and more than 1.0 cd/m² at 10 m, with diminishing returns above that, while lamp type made little measurable difference. The critical task after dark is reading an approaching person, which depends on light on faces at walking distance.

Analysis: what fails after 9 pm

Figure 1. The same street block by day and after 9 pm: lit and open frontage (amber) versus shuttered edges (indicative scenario on an OpenStreetMap figure-ground, Laxmi Road, Pune).
Source: Author's drawing based on OpenStreetMap data (© OpenStreetMap contributors, ODbL), 2026. The night pattern is indicative, not a survey.

Figure 1 draws the same block of Laxmi Road in Pune's old city on an OpenStreetMap figure-ground, once by day and once after 9 pm. The night pattern is an indicative scenario rather than a survey. By day, about three-quarters of the street-facing edge is drawn as open frontage, and pedestrians are spread along the whole block. At night, only about a third of that edge stays lit, gathered around a chemist, a food stall and a junction, and the islands of light are separated by long shuttered stretches. The figure-ground does not change; what changes is which edges still return light and attention to the footpath. A corridor can pass a pole-count audit and still feel empty, because the darkness that matters sits between the islands.

Figure 2. Pole-top road lighting versus pedestrian-scale light and lit frontage (paired sections at one scale, shared ground datum).
Source: Author's drawing (indicative sections; not a lighting calculation).

Figure 2 cuts two sections of the same width on a shared ground line. On the left, tall pole-top lamps, of the kind used on arterial roads, throw their light over the carriageway. Compound walls line both edges, and the footpath falls into a band of shadow beside them. On the right, lamps at about 4.5 m, spill from open shopfronts and a vendor's lamp light the walk itself, and the dotted line marks the zone where faces are seen at 1.5–1.7 m. The sections are not lighting calculations; they show a design choice that the Delhi visibility findings and the face-recognition studies both point to: light placed for vehicles and edges closed to the street together produce a walk that is technically lit and socially empty.

The NIGHT-5 diagnostic

Figure 3. NIGHT-5: five field tests for the night-time public realm.
Source: Author's framework drawing (indicative sketches, not to scale).

The five tests below refine the audit parameters into questions about street form that can be answered on a single walk after 9 pm (Figure 3).

1. Lit active frontage. What share of the street edge has lit, open or glazed ground floors once the main shops close? Late-open chemists, eateries, residential windows and lit lobbies count; a shutter or a blank compound wall does not. This test carries Jacobs's (1961) watched street into the night hours.

2. Continuous lit walk. Is there a walkable footpath that is lit from end to end between the places people actually travel: a stop, a gate, a lit edge? The audit should mark every dark gap rather than average the lux level along the street, since one gap can deter a whole trip (Safetipin, 2019).

3. Eyes and late uses. Are there chai stalls, dhabas, chemists, paan shops or vendors that stay open, and are women and families among the people present? Informal late uses are often removed in the name of order, yet the Delhi data link roadside commerce and vendors with better visibility scores (Safetipin, 2019).

4. Lit, overlooked wait points. Are bus stops, auto stands and metro exits lit, and are they within sight of an open use? The audit treats access to public transport as part of safety (Viswanath & Basu, 2015), and waiting is when people are most exposed.

5. Faces, not only floors. Does light reach faces at walking height, with low glare, so that a face can be read about 10 m ahead? This calls for lower, shielded fittings and light from frontages, rather than ever brighter pole-top lamps on the carriageway (Fotios et al., 2015).

Applying the diagnostic

The tests are designed for a two-person walk on a weekday after 9 pm, using a phone for photographs and a simple tally sheet. Each test can be scored 0–3 along a block, in line with the audit scale, and mapped onto a figure-ground like Figure 1. The scores also sort interventions by cost. Repairing lamps and lighting wait points are quick public works; licensing late vending in defined spots, asking new buildings for lit frontage instead of boundary walls, and adding pedestrian-scale lamps take longer but last. Street-lighting contracts could then report these indicators, not only lamp uptime.

Limits

The evidence base has clear limits. The lighting and crime review draws on studies from high-income countries, and whether its effects transfer to Indian arterials with mixed traffic and informal commerce is untested (Welsh & Farrington, 2008). The face-recognition results come from controlled laboratory conditions in the United Kingdom (Fotios et al., 2015). Safety audit data measure perceived and observed conditions rather than recorded crime, and crowdsourced coverage can vary between neighbourhoods (Viswanath & Basu, 2015). Figures 1 and 2 are indicative drawings. NIGHT-5 should therefore be read as a structured field diagnostic for comparing streets and priorities, not as a validated safety index.

Conclusion

A street that is full at noon and empty at night has rarely lost its lamps; more often it has lost its open edges, its continuous walk and its reasons to linger. The evidence suggests that light works partly by signalling care, and that blind edges are harder to fix than broken poles. The practical implication is to appraise the night-time street as a piece of urban form: keep frontages lit and open, close the dark gaps, protect late informal uses, light the places where people wait, and light faces rather than only carriageways. The main limitation is that the diagnostic is still to be tested against recorded outcomes in Indian cities, and that test is the next step.

Notes

Method and sources. This article reviews peer-reviewed studies, a city-wide safety assessment report and two key books, consulted in October 2026, and applies them to drawings of a Pune street. Data. Building footprints from OpenStreetMap (© OpenStreetMap contributors, ODbL), extract cached in 2026. Interests. No competing interests. Preparation note. AI tools assisted with research, drafting and drawings; the author reviewed the sources and text.

References

Fotios, S., Yang, B., & Cheal, C. (2015). Effects of outdoor lighting on judgements of emotion and gaze direction. Lighting Research & Technology, 47(3), 301–315. https://doi.org/10.1177/1477153513510311

Gehl, J. (2010). Cities for people. Island Press.

Jacobs, J. (1961). The death and life of great American cities. Random House.

Safetipin. (2019). Delhi: A safety assessment report [Report sponsored by the Department of Women and Child Development, Government of NCT of Delhi]. https://safetipin.com/wp-content/uploads/2019/12/delhi_a_safety_assessment_report_2019.pdf

Viswanath, K., & Basu, A. K. (2015). SafetiPin: An innovative mobile app to collect data on women's safety in Indian cities. Gender & Development, 23(1), 45–60. https://doi.org/10.1080/13552074.2015.1013669

Welsh, B. C., & Farrington, D. P. (2008). Effects of improved street lighting on crime. Campbell Systematic Reviews, 4(1). https://doi.org/10.4073/csr.2008.13