Mapping of towns/villages

How do we distinguish towns from villages?

In the Indian wiki page we have a proposal : https://wiki.openstreetmap.org/wiki/India/Places
There has been a talk-IN discussion in 2012 and 2016:
http://lists.openstreetmap.org/pipermail/talk-in/2012-May/thread.html
https://lists.openstreetmap.org/pipermail/talk-in/2016-August/thread.html#2654
Most mappers opposed the wiki proposal but there was no clear result.
I think this proposal is not practicable and I do not see that India needs a special definition of towns and villages.

Then census data make the following definition :
To be called a town, a place should meet all the following criteria:
1)Population exceeds 5,000
2)At least 75% of male working population is employed outside the
agricultural sector
3)Minimum population density of 400 persons per km
My question : How does a normal mapper get the information for 2) and 3)??

So we should look for the common mapping worldwide.

The the opinions is as follows:

  • „Use place=town to identify an important urban centre that is larger than a place=village, smaller than a place=city, and not a place=suburb. Towns normally have a good range of shops and facilities which are used by people from nearby villages. „
  • towns : population= 10 000 to 100 000
  • towns : population = 5 000 to 10 000 and important center for the region

Where can we find the population data of towns?

**So I make the following proposal :

  • towns : population= 10 000 to 100 000
  • towns : population = about 5 000 to 10 000 and the place is a district or subdistrict headquarter.
    **
    All districts are mapped with the headquarter and nearly all subdistricts.

What is your opinion?

Its simple and practically makes sense. I’m for it!

I think that this makes good sense as far as the population classifications are concerned.

But for subdistricts and their HQs - is the term ‘subdistrict’ applied consistently between states? Do they have similar importance in each state?

For subdistrict see here : https://wiki.openstreetmap.org/wiki/Subdistricts_in_India

Yes , I think they are applied consistently and have similiar importance.

The subdistrict headquarters normally have a good range of shops and facilities which are used by people from nearby villages.

Rural -Urban Classification for Census - 2021

http://censusindia.gov.in/2021-Census/Circular-2-2021.pdf

Analysis of the town definitions in the census 2011 in Rajasthan

Definition of towns in census 2011 :
(a) Statutory Towns:All places with a municipality, corporation, cantonment board or notified town area committee, etc. are known as statutory towns.
(b) Census Towns:All other places satisfying the following three criteria simultaneously are treated as Census Towns.
i) A minimum population of 5,000;
ii) At least 75 per cent of male working population
engaged in non-agricultural pursuits; and
iii) A density of population of at least 400 per sq. km. (1,000 per sq. Mile)

There are
185 statutory towns
112 census towns
297 total number of towns

44672 total number of villages
120 villages with population > 10 000 ( upto 26471)!!!

There are 57 towns with a population < 10000 (2525 to 9999).
All these towns are census towns not statutory towns.
That means, all the administrations (listed in the wiki-proposal) are not used for the town definition!

10 census towns have a population < 5000.
I do not know how they manage to fulfill the rule (i)!!! I think, they should be mapped as villages in OSM.

So there are 47 census towns with : 5000< =population<=10000
7 census towns : - unknown reason for the declaration
- place without a center but large area
- small village with nearby important industrial complex, airport…
- small village with touristic hotspot
19 census town are the headquarter of a subdistrict
9 census towns are suburbs of a big city and should be mapped in OSM as suburbs.
12 census towns fulfill the rules I, II, III

My conclusions:

  1. The administrations of the wiki-proposal are not needed because all these statutory towns have a population > 10 000.
  2. places with a population < 5000 should in any case be mapped as villages
  3. May be the situation is different in other states.
    So I propose to adopt my proposal till the census 2021 is done. Then we may decide, whether there are additional census towns which should be mapped in OSM as towns.

My opinion, for what it is worth:
*) we should not be concocting our own norms or definitions or criteria, but only follow those that are in place, from sources of authority
*) even if we were to set up our own definitions, they should be practical and politically tenable. Definitions based on people’s gender are quite unacceptable, these days, for one example.

If we do have to define a “town” for ourselves, even if I do not like the idea at all, important criteria should be the availability of services, such as schools of various levels, hospital, transportation infrastructure, and more such.

My preference would be to only publish what is beyond discussion:

  • this is an inhabited place
  • its name is xxx
  • population=9999 , if possible with a source reference and a date

You wrote :
„My opinion, for what it is worth:
*) we should not be concocting our own norms or definitions or criteria, but only follow those that are in place, from sources of authority“

It is a fact that OSM has a definite opinion about place mapping (see the wiki). So we do not have to follow exactly the definitions of authorities in a country. It is more important that the place mapping wordwide is similiar in OSM. Of course every country can look for a reasonable, additional rules.

You wrote :
„*) even if we were to set up our own definitions, they should be practical and politically tenable. Definitions based on people’s gender are quite unacceptable, these days, for one example.“

See my last post : My proposal comes in Rajasthan to quite similiar results as census 2011. There are only 12 census towns of 297 towns, where census comes to a different result. In the other cases (16 places) the census decisions are completely wrong, I think. And it is a bit strange to declare a place with population > 10000 as a village.

You wrote:
„If we do have to define a “town” for ourselves, even if I do not like the idea at all, important criteria should be the availability of services, such as schools of various levels, hospital, transportation infrastructure, hospitals.“

As far as I can see, the headquarters of subdistricts are the places with the infrastructure you mention. The progress of the India map in OSM is remarkable in the last years. But still the number of mappers with very good mapping practise is still small.
An example: While mapping subdistricts I saw hundreds of place nodes which were imported 12 years ago by AND. But often nobody looked for the position errors (up to 50km!), nobody mapped the exact population and the imported nodes are completely untouched.
That means, we need a practicable and reasonable rule what is a town and what is a village.
When the census 2021 is done, we have reliable data to correct the census towns (if wanted) and that is an easy task.

You wrote :
„My preference would be to only publish what is beyond discussion:

  • this is an inhabited place
  • its name is xxx
  • population=9999 , if possible with a source reference and a date“

That is not possible. If you want to map the name and the population of a place, you have to put a place tag and also decide whether it is a hamlet, village,…… city. Then you can add name and population.
By the way as far as I know, nobody added a date-tag to the population. Nevertheless that is a really usefull information.

Will this help ?

http://censusindia.gov.in/2011census/Listofvillagesandtowns.aspx

No, that does not help. The list gives only the census code number of towns and villages.

For my analysis of the situation in Rajasthan I used the original list here:
http://censusindia.gov.in/2011census/dchb/DCHB.html
In the column „Town amenities“ and „village amenities“ we have list of all towns and villages with their population and declaration (like the wiki-proposal). Of special interest are the census towns (CT).

As a result of this discussion we could do it like this:

  1. mapping of towns and villages according to my proposal
  2. control of the census towns in the mentioned list with 5000<population<1000 :
    a) if they are suburbs of a big city: mapping with place=suburb
    b) not a suburb : correction from village to town
  3. control of villages in the list with population>10000 : upgrade from village to town (to be discussed)

As a result we have the following difference to census declarations:
a) villages with population>10000 will be towns
b) suburb places will not be towns
c) census towns with population<5000 will be villages

If we come to this result, mappers are necessary who do the corrections. I do not see that a normal mapper does this controling.

+1 to the proposal.

There are in Rajasthan 120 census villages with population>10 000
Worldwide in OSM places with a population>10000 are mapped as towns. 34 of the 120 census villages are also subdistrict headquarters. I checked some of these places and I found that there are market areas, secondary schools, bank branches, health facilities, shops…
It seems they are declared census villages because more than 75% of the male population are still agricultural workers. For me these are all towns.

17 places are not subdistrict headquarters, they have a population<10000 and are census towns but not the suburb of a big city.
From time to time we should check these cases with the census lists and make corrections (upgrade to town).

I suppose that the normal mappers will still choice town/village without the mentioned checks.

So the best is

  • when in each state there is a mapper who checks the census lists
    (http://censusindia.gov.in/2011census/dchb/DCHB.html) for population values and the
    mentioned town/villages rules and adds a note on the place tag

  • when for each state a list of OSM towns are published in the state wiki as a reference.

For Rajasthan I can start with the check, create a list in the wiki and publish the exact rules for the village/town decision in the wiki.

I posted the latest version of my proposal here: https://wiki.openstreetmap.org/wiki/Talk:India/Places

I controlled all towns in Rajasthan according to these rules. Here is the resulting list of towns:

https://wiki.openstreetmap.org/wiki/Rajasthan

I would agree with this proposal on that page but I’m currently having some trouble logging in on the wiki pages. So just count this post please. Thanks,