Pune tanker dispatch
Pune's corporation runs the tanker fleet itself and publishes what it sent: one spreadsheet per filling point per working day, one row per tanker. This is neither a booking system nor a household survey, so it does not show what anyone asked for or what anyone paid - it shows what the fleet actually did, and whether each trip was on a rota or a response.
Most of this water was not planned for
Pune is one of the few Indian cities where the corporation runs the tanker fleet itself and publishes the dispatch record - a spreadsheet per filling point per working day, one row per tanker sent. It is not a booking system and not a survey: it is what the fleet actually did. Each row records whether the trip was scheduled, meaning a known supply gap being serviced on a rota, or on demand, meaning somebody rang. 58.4% of trips are on demand. A city meeting known shortfalls on a rota would show the opposite ratio.
Computed only over rows from a layout that HAS an on-demand column. The Bund Garden sheets do not carry one at all, so including them would push the on-demand share toward zero for a reason unrelated to demand.
2,083 deliveries to an area that is no longer in the corporation
2,083 deliveries - 6.9% of everything with a ward attached - are booked in PMC's own register to NAGAR PARISHAD (Uruli Devachi / Fursungi) rather than to a prabhag. The tankers did not stop at the new boundary. Kept here as its own line rather than folded into a ward number, because it is not a ward.
Where the tankers fill
Two points carry most of the load. The mean is over the days a point could be dated, not over the whole window, which is why a point with few reporting days can show a high daily mean.
Deliveries by prabhag: a partial attribution
The prabhag column is filled on 52.8% of rows, so this is not a ward ranking and the order below should not be read as which prabhag is worst served. A ward can be low here because its deliveries were recorded without a prabhag. Showing the top 12 of 32 that appear at all.
What this data does not tell you
39,332 of 57,370 rows cannot be dated. Registers with no date anywhere - not in the title, not in the sheet. PMC's upload timestamp is deliberately NOT used as a substitute: it is when the file was posted, not the day the tankers ran. So there is no usable daily demand series here, only 104 days that could be dated at all, and every per-day figure on this page is computed over that subset rather than the whole register.
Coverage starts when PMC began publishing these registers, not when tanker supply began. There is no earlier archive on this endpoint.
No volumes. The register counts tanker trips, not litres. Tanker capacity is not recorded on the row, so the water this represents cannot be totalled without assuming a load size that PMC does not publish.
No recipients, by choice. The source rows carry society names, street addresses and phone numbers. None of that is republished, and it is not in the artifact behind this page. That is deliberately stricter than the treatment of a bulk-sales ledger, where the buyers are companies purchasing at a published tariff: these recipients are private housing societies, and naming one alongside its address publishes where identifiable residents do not have water.
Only what PMC itself sent. Pune has a private tanker market alongside this one and no public record of it exists, so this is the floor on tanker dependence rather than the total.
11 of 411 published registers could not be parsed and are excluded. 1,487 trip cells held a misaligned vehicle number rather than a count and were rejected rather than summed; summing them naively would have reported millions of trips against 57,370 rows.
Source: PMC daily tanker delivery registers · fetched 2026-08-17 · aggregated on build, published as counts only