A national estimate is not a map.
Almost every state publishes a street-dog estimate. Two publish a sterilisation rate. None publishes either ward by ward, which is the only scale the work happens at. We build that record, method attached.
Twenty-eight states, eight union territories. Two have a coverage figure.
Coloured by what has been published, not by how much work is happening there: by whether a figure exists that somebody outside the programme could check. Almost every state estimates how many street dogs it has. Almost none says how many have been reached.
- Delhi
- Goa
- Andaman & Nicobar Islands
- Andhra Pradesh
- Arunachal Pradesh
- Assam
- Bihar
- Chandigarh
- Chhattisgarh
- Dadra & Nagar Haveli and Daman & Diu
- Gujarat
- Haryana
- Himachal Pradesh
- Jammu & Kashmir
- Jharkhand
- Karnataka
- Kerala
- Ladakh
- Lakshadweep
- Madhya Pradesh
- Maharashtra
- Manipur
- Meghalaya
- Mizoram
- Nagaland
- Odisha
- Puducherry
- Punjab
- Rajasthan
- Sikkim
- Telangana
- Tripura
- Uttar Pradesh
- Uttarakhand
- West Bengal
- Tamil Nadu
- 2 coverage published
- 33 population only
- 1 nothing published
Built from the published sources listed on sources, and drawn from the same classification the map's coverage layer uses. A state moves colour when somebody publishes a figure for it, never when we estimate one.
It is a measurement problem, not an effort problem.
The ABC (Dogs) Rules require every municipality to run a sterilisation programme. Rescues and municipal teams do that work, often for decades. What almost nobody has is a denominator: a count of the animals in an area, so a sterilisation figure can be a rate rather than a number of surgeries.
Chennai is a fair example. Long-running ABC work by named organisations, widely cited coverage estimates, and no official ward-level figure published by the corporation. That is not a criticism of the people doing the work. It is a gap in the record, and the gap is what we are here for.
What gets counted
Surgeries. A real number, kept carefully, reported upward, and on its own it cannot tell you whether a street is covered.
What has to be counted
The animals themselves, inside a boundary somebody recognises. Then a surgery count becomes coverage, and coverage can be argued with, funded and checked next year.
Five rules, and each one is a decision in the software.
Not values on a wall. Every one of these is a place where the easier build was available and we did not take it, and the cost of that choice shows up underneath each rule.
An area with no records is not an area with no animals.
It is an area nobody has surveyed. Those are opposite findings, and a map that shades them the same colour tells a funder an unvisited ward is a quiet one.
- A published figure
- Half a figure
- Nobody has looked, drawn outside the scale
A rate has to say what it is a rate of.
Every sterilisation percentage on this site is shown twice: of the animals somebody actually checked, and of every animal on record with the unchecked counted against.
Kindersterilised / checkedHoneststerilised / all on record“Nobody checked” is a real answer.
A yes-or-no field cannot say that nobody has looked, so it quietly files every unexamined animal alongside the confirmed ones. Ours has three states, and the third is the default.
SterilisedNot sterilisedNobody checkedNothing reaches an organisation's dashboard on its own.
A volunteer's report waits until that organisation files it. A community sighting has to be claimed. Friction is the point: numbers nobody accepted responsibility for are numbers nobody can defend.
Sighting↓
waits↓Claimed by a team↓On the registerEvery figure names its source.
Boundaries carry the dataset they came from and its licence. Costs cite the municipal tender or the notified ceiling. Where a number does not exist we say so rather than estimating one.
45%Delhi ABC coverage, 2023. 2022-23 community-dog population survey cited in Delhi rabies-elimination reporting.
What we claim, and what we do not.
What our numbers are
- Animals recorded in StrayPaw, in a place, on a date.
- A count inside published municipal boundaries, so a ward, a district and a state each have a number that came from somewhere.
- One animal followed across the years, rather than a new record every time somebody new sees it.
What they are not
- An estimated population. In most of the country the honest figure today is zero surveyed, and we would rather show that than a modelled number that answers a question nobody asked.
- A rescue service. When an animal needs help the people who go out are the organisations already doing it.
- A ranking of anybody's effort. A blank state is a blank record, not a lazy one.
Three people should be able to open the same page.
None of this requires anything clever. It requires the records to be kept properly, in one place, by the people already doing the work, and published in a form somebody can argue with. That is the whole plan.
Opens their ward and sees what is known about it: how many animals are on the record, how many have been checked, and what has not been visited since.
Sees the same numbers the organisation sees, drawn against boundaries both of them recognise, so a review is about the work rather than about whose spreadsheet is right.
Downloads the table, follows every figure to its source, and finds the arithmetic holds.
The longer argument What is known and missing Every figure we cite Built in India by Jay.