Re: [datameet] Re: Correct CRS for datameet maps

[Ashim Kapoor]

Dear Sharad,

My apologies for the delay in responding.(The week was blissfully busy and
I lost track of responding to this email). My response is inline.

On Thu, Apr 18, 2019 at 11:09 PM Sharad Lele  wrote:

> Devdatta's summary was short and sweet, and practical. To add to it:
>
> 1. The NNRMS CRS suggested seems to be in Lambert Conformal Conic
> projection (with WGS84 ellipsoid and datum). An alternative to it is the
> UTM projection system, which keeps longitudes parallel, and gives
> reasonably accurate area estimate. The particular UTM zone one may use
> varies by which part of India one is in.
> 2. In common parlance, 'unprojected' means having no Coordinate Reference
> System at all. In GIS parlance, 'unprojected' may be understood as
> 'geographic' CRS, that is, in lat-long rather than in metres. I presume
> Ashim means 'having no CRS' at all. Whether one's map needs a CRS or not
> depends upon what use one wants to put it to (and hence how accurate it
> must be) and its size. A map of a 60'x40' piece of land does not require a
> CRS if the purpose is house construction. It may require a CRS if one is
> setting up a telescope on it which will do interferometry with another
> telescope 500km away. A map of even a layout or a village or small town may
> not need a CRS if one wants approximate locations/directions to navigate
> visually (old style). It will need a CRS if one wants to use a GPS to
> navigate.
> If, Ashim, you mean 'geographic' vs 'projected' system, the answer is
> different: geographic is useful for representation, easy to understand
> (lat-long concept), but not to be used for calculating distances and areas.
>

I think when the data is "unprojected", it simply plots the points in a
plane based on the co-ordinates( long / lat) . When it is projected it uses
a transformation to go from points ( long / lat ) -> 2D map.


> Hope this helps. Devdatta and others: please correct/supplement.
> Sharad
>
> On Tuesday, April 16, 2019 at 3:38:08 PM UTC+5:30, Ashim wrote:
>>
>> Dear All,
>>
>> I have a small query on similar lines. When do we use  UNPROJECTED maps ?
>> I understand that to go from 3D to 2D we need a projection. When is it
>> reasonable to use an unprojected map?
>>
>> Many thanks,
>> Ashim
>>
>> On Tue, Apr 16, 2019 at 12:12 PM Ashim Kapoor  wrote:
>>
>>> Dear Devdatta and Nikhil,
>>>
>>> I attended a R GIS school where the teachers were from France and they
>>> were using CRS 2154. They told us that they were not familiar with the CRS
>>> for Indian maps.
>>>
>>> I used 2154 thinking that it would not make much  of a difference, but I
>>> was wrong. My apologies for the confusion. When I used that on Indian (
>>> datameet ) shapefiles,I got this : map1.png ( see attachment )
>>>
>>> I have also attached the R code to create these maps from the Datameet
>>> District level ( 2011 ) shapefiles. I have created map2.png and map3.png
>>> from crs = 3857 / 7755 respectively.
>>>
>>> I also found this :
>>> https://epsg.io/?q=india
>>>
>>> Which is confusing because many CRS are there for India. Why do we have
>>> so many choices of CRS for India ?
>>>
>>> I guess for the time being I will use 7755.
>>>
>>> Many thanks to Devdatta and Nikhil for their help.
>>> Best Regards,
>>> Ashim
>>>
>>>
>>>
>>> On Tue, Apr 16, 2019 at 11:42 AM Nikhil VJ  wrote:
>>>
 Hi Ashim,

 Pls provide the exact link of what you're referring - this stuff is
 never as much on the top of people's minds as we assume it to be.
 Do try loading your shapefiles on other tools and check there :
 mapshaper.org is a site that does a lot of quick, cool things with
 shapefiles; QGIS is a software that does everything with shapefiles. Both
 are free and open source.

 The latter will help you re-save the shapefile into any CRS you want.
 Right-click the layer > Save As (or "Export") and remember to choose your
 preferred CRS in the dropdown.

 Can you give background on what this CRS : *2154* : is and why you
 want to transform to that? Because I've only come across two main CRS's
 both being under the "WSG 84" category :
 - EPSG 4326 : this makes everything in latitude longitude
 - EPSG 3857 : this makes everything in meters from the equator (I
 think) and we need to get the data in this format when we want to do things
 in physical distance terms like making a 1km distance buffer or measuring
 areas

 I haven't learned GIS stuff from theory, I just use it, so don't know
 more details about CRS. I do understand that the dizzying multitude of
 CRS's out there are so because apparently GIS folks like to re-orient the
 center of the world (geometrically speaking) to where their data is to
 ensure least distortion of their shapes in the rendering. When I come
 across anything that's in a non-standard CRS, my first move is to transform
 it to either lat-longs (4326) or meters (3857).

 Regards
 

Re: [datameet] Re: Correct CRS for datameet maps

[शरच्चंद्र लेले]

Nice! I did not know this!


On 19-04-2019 06:46, Naraina Damle wrote:

I would like to add my two cents to this discussion.

You need to read, study, learn and practice this stuff about 
Coordinate System and projections.


You can always land up in surprise\ing situations.

One thing I would like to mention here is to Split the thing in Two 
Aspects : DATUM and PROJECTION and understand the difference between 
the two.



Simply stated just saying "WGS84" is not complete. It is only talking 
about Datum and not projections.


DATUM DECIDES WHAT IS THE PRECISE LATITUDE AND LONGITUDE OF  A PLACE.

As per WGS84 the ZERO Longitude is NOT at Greenwhich line  it is a 
little away.


https://gis.stackexchange.com/questions/88428/why-is-the-gps-reference-meridian-100m-to-the-east-of-the-prime-meridian 



One needs to Know and use Datum even if you want to make a Map on a 
Actually Sphere / Globe. But in this case you do not need projection.


PROJECTION DECIDES HOW YOU ARE FLATTENING THE SPHERE/ PART OF IT TO 
SEE IT ON A FLAT MAP.


So in case of a Flat map You need to be careful about DATUM as well as 
PROJECTION.


Datum is required because there were historical maps made with Datum 
other than WGS84.


.. This is just a simplified way but it is more complex. I am a 
permanent learner.


Regards
Naraina Damle



On Thursday, April 18, 2019 at 11:08:57 PM UTC+5:30, Sharad Lele wrote:

Devdatta's summary was short and sweet, and practical. To add to it:

1. The NNRMS CRS suggested seems to be in Lambert Conformal Conic
projection (with WGS84 ellipsoid and datum). An alternative to it
is the UTM projection system, which keeps longitudes parallel, and
gives reasonably accurate area estimate. The particular UTM zone
one may use varies by which part of India one is in.
2. In common parlance, 'unprojected' means having no Coordinate
Reference System at all. In GIS parlance, 'unprojected' may be
understood as 'geographic' CRS, that is, in lat-long rather than
in metres. I presume Ashim means 'having no CRS' at all. Whether
one's map needs a CRS or not depends upon what use one wants to
put it to (and hence how accurate it must be) and its size. A map
of a 60'x40' piece of land does not require a CRS if the purpose
is house construction. It may require a CRS if one is setting up a
telescope on it which will do interferometry with another
telescope 500km away. A map of even a layout or a village or small
town may not need a CRS if one wants approximate
locations/directions to navigate visually (old style). It will
need a CRS if one wants to use a GPS to navigate.
If, Ashim, you mean 'geographic' vs 'projected' system, the answer
is different: geographic is useful for representation, easy to
understand (lat-long concept), but not to be used for calculating
distances and areas.

Hope this helps. Devdatta and others: please correct/supplement.
Sharad

On Tuesday, April 16, 2019 at 3:38:08 PM UTC+5:30, Ashim wrote:

Dear All,

I have a small query on similar lines. When do we use 
UNPROJECTED maps ? I understand that to go from 3D to 2D we
need a projection. When is it reasonable to use an unprojected
map?

Many thanks,
Ashim

On Tue, Apr 16, 2019 at 12:12 PM Ashim Kapoor
 wrote:

Dear Devdatta and Nikhil,

I attended a R GIS school where the teachers were from
France and they were using CRS 2154. They told us that
they were not familiar with the CRS for Indian maps.

I used 2154 thinking that it would not make much  of a
difference, but I was wrong. My apologies for the
confusion. When I used that on Indian ( datameet )
shapefiles,I got this : map1.png ( see attachment )

I have also attached the R code to create these maps from
the Datameet District level ( 2011 ) shapefiles. I have
created map2.png and map3.png from crs = 3857 / 7755
respectively.

I also found this :
https://epsg.io/?q=india

Which is confusing because many CRS are there for India.
Why do we have so many choices of CRS for India ?

I guess for the time being I will use 7755.

Many thanks to Devdatta and Nikhil for their help.
Best Regards,
Ashim



On Tue, Apr 16, 2019 at 11:42 AM Nikhil VJ
 wrote:

Hi Ashim,

Pls provide the exact link of what you're referring -
this stuff is never as much on the top of people's
minds as we assume it to be.
Do try loading your shapefiles on other tools and
check there : mapshaper.org  is
a site that does a lot of quick, cool things with

Re: [datameet] Re: Correct CRS for datameet maps

[Naraina Damle]

I would like to add my two cents to this discussion. 

You need to read, study, learn and practice this stuff about Coordinate 
System and projections. 

You can always land up in surprise\ing situations. 

One thing I would like to mention here is to Split the thing in Two Aspects 
: DATUM and PROJECTION and understand the difference between the two. 


Simply stated just saying "WGS84" is not complete. It is only talking about 
Datum and not projections. 

DATUM DECIDES WHAT IS THE PRECISE LATITUDE AND LONGITUDE OF  A PLACE. 

As per WGS84 the ZERO Longitude is NOT at Greenwhich line  it is a little 
away. 

https://gis.stackexchange.com/questions/88428/why-is-the-gps-reference-meridian-100m-to-the-east-of-the-prime-meridian
 

One needs to Know and use Datum even if you want to make a Map on a 
Actually Sphere / Globe. But in this case you do not need projection. 

PROJECTION DECIDES HOW YOU ARE FLATTENING THE SPHERE/ PART OF IT TO SEE IT 
ON A FLAT MAP. 

So in case of a Flat map You need to be careful about DATUM as well as 
PROJECTION. 

Datum is required because there were historical maps made with Datum other 
than WGS84.

.. This is just a simplified way but it is more complex. I am a 
permanent learner.

Regards
Naraina Damle



On Thursday, April 18, 2019 at 11:08:57 PM UTC+5:30, Sharad Lele wrote:
>
> Devdatta's summary was short and sweet, and practical. To add to it: 
>
> 1. The NNRMS CRS suggested seems to be in Lambert Conformal Conic 
> projection (with WGS84 ellipsoid and datum). An alternative to it is the 
> UTM projection system, which keeps longitudes parallel, and gives 
> reasonably accurate area estimate. The particular UTM zone one may use 
> varies by which part of India one is in.
> 2. In common parlance, 'unprojected' means having no Coordinate Reference 
> System at all. In GIS parlance, 'unprojected' may be understood as 
> 'geographic' CRS, that is, in lat-long rather than in metres. I presume 
> Ashim means 'having no CRS' at all. Whether one's map needs a CRS or not 
> depends upon what use one wants to put it to (and hence how accurate it 
> must be) and its size. A map of a 60'x40' piece of land does not require a 
> CRS if the purpose is house construction. It may require a CRS if one is 
> setting up a telescope on it which will do interferometry with another 
> telescope 500km away. A map of even a layout or a village or small town may 
> not need a CRS if one wants approximate locations/directions to navigate 
> visually (old style). It will need a CRS if one wants to use a GPS to 
> navigate.
> If, Ashim, you mean 'geographic' vs 'projected' system, the answer is 
> different: geographic is useful for representation, easy to understand 
> (lat-long concept), but not to be used for calculating distances and areas.
>
> Hope this helps. Devdatta and others: please correct/supplement.
> Sharad
>
> On Tuesday, April 16, 2019 at 3:38:08 PM UTC+5:30, Ashim wrote:
>>
>> Dear All,
>>
>> I have a small query on similar lines. When do we use  UNPROJECTED maps ? 
>> I understand that to go from 3D to 2D we need a projection. When is it 
>> reasonable to use an unprojected map?
>>
>> Many thanks,
>> Ashim
>>
>> On Tue, Apr 16, 2019 at 12:12 PM Ashim Kapoor  wrote:
>>
>>> Dear Devdatta and Nikhil,
>>>
>>> I attended a R GIS school where the teachers were from France and they 
>>> were using CRS 2154. They told us that they were not familiar with the CRS 
>>> for Indian maps.
>>>
>>> I used 2154 thinking that it would not make much  of a difference, but I 
>>> was wrong. My apologies for the confusion. When I used that on Indian ( 
>>> datameet ) shapefiles,I got this : map1.png ( see attachment ) 
>>>
>>> I have also attached the R code to create these maps from the Datameet 
>>> District level ( 2011 ) shapefiles. I have created map2.png and map3.png 
>>> from crs = 3857 / 7755 respectively.
>>>
>>> I also found this : 
>>> https://epsg.io/?q=india
>>>
>>> Which is confusing because many CRS are there for India. Why do we have 
>>> so many choices of CRS for India ? 
>>>
>>> I guess for the time being I will use 7755.
>>>
>>> Many thanks to Devdatta and Nikhil for their help.
>>> Best Regards,
>>> Ashim
>>>
>>>
>>>
>>> On Tue, Apr 16, 2019 at 11:42 AM Nikhil VJ  wrote:
>>>
 Hi Ashim,

 Pls provide the exact link of what you're referring - this stuff is 
 never as much on the top of people's minds as we assume it to be.
 Do try loading your shapefiles on other tools and check there : 
 mapshaper.org is a site that does a lot of quick, cool things with 
 shapefiles; QGIS is a software that does everything with shapefiles. Both 
 are free and open source.

 The latter will help you re-save the shapefile into any CRS you want. 
 Right-click the layer > Save As (or "Export") and remember to choose your 
 preferred CRS in the dropdown.

 Can you give background on what this CRS : *2154* : is and why you 
 want to 

Re: [datameet] Re: Correct CRS for datameet maps

[Sharad Lele]

Devdatta's summary was short and sweet, and practical. To add to it: 

1. The NNRMS CRS suggested seems to be in Lambert Conformal Conic 
projection (with WGS84 ellipsoid and datum). An alternative to it is the 
UTM projection system, which keeps longitudes parallel, and gives 
reasonably accurate area estimate. The particular UTM zone one may use 
varies by which part of India one is in.
2. In common parlance, 'unprojected' means having no Coordinate Reference 
System at all. In GIS parlance, 'unprojected' may be understood as 
'geographic' CRS, that is, in lat-long rather than in metres. I presume 
Ashim means 'having no CRS' at all. Whether one's map needs a CRS or not 
depends upon what use one wants to put it to (and hence how accurate it 
must be) and its size. A map of a 60'x40' piece of land does not require a 
CRS if the purpose is house construction. It may require a CRS if one is 
setting up a telescope on it which will do interferometry with another 
telescope 500km away. A map of even a layout or a village or small town may 
not need a CRS if one wants approximate locations/directions to navigate 
visually (old style). It will need a CRS if one wants to use a GPS to 
navigate.
If, Ashim, you mean 'geographic' vs 'projected' system, the answer is 
different: geographic is useful for representation, easy to understand 
(lat-long concept), but not to be used for calculating distances and areas.

Hope this helps. Devdatta and others: please correct/supplement.
Sharad

On Tuesday, April 16, 2019 at 3:38:08 PM UTC+5:30, Ashim wrote:
>
> Dear All,
>
> I have a small query on similar lines. When do we use  UNPROJECTED maps ? 
> I understand that to go from 3D to 2D we need a projection. When is it 
> reasonable to use an unprojected map?
>
> Many thanks,
> Ashim
>
> On Tue, Apr 16, 2019 at 12:12 PM Ashim Kapoor  > wrote:
>
>> Dear Devdatta and Nikhil,
>>
>> I attended a R GIS school where the teachers were from France and they 
>> were using CRS 2154. They told us that they were not familiar with the CRS 
>> for Indian maps.
>>
>> I used 2154 thinking that it would not make much  of a difference, but I 
>> was wrong. My apologies for the confusion. When I used that on Indian ( 
>> datameet ) shapefiles,I got this : map1.png ( see attachment ) 
>>
>> I have also attached the R code to create these maps from the Datameet 
>> District level ( 2011 ) shapefiles. I have created map2.png and map3.png 
>> from crs = 3857 / 7755 respectively.
>>
>> I also found this : 
>> https://epsg.io/?q=india
>>
>> Which is confusing because many CRS are there for India. Why do we have 
>> so many choices of CRS for India ? 
>>
>> I guess for the time being I will use 7755.
>>
>> Many thanks to Devdatta and Nikhil for their help.
>> Best Regards,
>> Ashim
>>
>>
>>
>> On Tue, Apr 16, 2019 at 11:42 AM Nikhil VJ > > wrote:
>>
>>> Hi Ashim,
>>>
>>> Pls provide the exact link of what you're referring - this stuff is 
>>> never as much on the top of people's minds as we assume it to be.
>>> Do try loading your shapefiles on other tools and check there : 
>>> mapshaper.org is a site that does a lot of quick, cool things with 
>>> shapefiles; QGIS is a software that does everything with shapefiles. Both 
>>> are free and open source.
>>>
>>> The latter will help you re-save the shapefile into any CRS you want. 
>>> Right-click the layer > Save As (or "Export") and remember to choose your 
>>> preferred CRS in the dropdown.
>>>
>>> Can you give background on what this CRS : *2154* : is and why you want 
>>> to transform to that? Because I've only come across two main CRS's both 
>>> being under the "WSG 84" category : 
>>> - EPSG 4326 : this makes everything in latitude longitude
>>> - EPSG 3857 : this makes everything in meters from the equator (I think) 
>>> and we need to get the data in this format when we want to do things in 
>>> physical distance terms like making a 1km distance buffer or measuring areas
>>>
>>> I haven't learned GIS stuff from theory, I just use it, so don't know 
>>> more details about CRS. I do understand that the dizzying multitude of 
>>> CRS's out there are so because apparently GIS folks like to re-orient the 
>>> center of the world (geometrically speaking) to where their data is to 
>>> ensure least distortion of their shapes in the rendering. When I come 
>>> across anything that's in a non-standard CRS, my first move is to transform 
>>> it to either lat-longs (4326) or meters (3857).
>>>
>>> Regards
>>> Nikhil VJ
>>>
>>>
>>> On Monday, April 15, 2019 at 1:57:20 PM UTC+5:30, Ashim wrote:

 Dear All,

 I am referring to 2011_Dist. * set of shapefile. (District level 
 shapefiles).

 When I read them as simple features in R, ( like this ) 

 map <- st_read("2011_Dist.shp") %>% st_as_sf()

 ( it's unprojected because it says 4326 in the epsg when I read the 
 above )

 plot(map$geometry)

 it looks OK.

 But when I do : 

Re: [datameet] Re: Correct CRS for datameet maps

[Nikhil VJ]

Hi Devdatta

Thanks a ton for the correction! I'll use EPSG:7755 for distance related
work here on out.

--
Cheers,
Nikhil VJ, Pune, India
http://nikhilvj.co.in


On Tue, Apr 16, 2019 at 11:56 AM Devdatta Tengshe 
wrote:

> HI Ashim, & others,
>
> Coordinate systems, and their area of applicability is an abstruse topic,
> and one that has many complicated things to consider. So what I'm saying
> here is a user-friendly summary.
>
>
>- The data in our Repository is in WGS-84 Lat/long, which has the EPSG
>code 4326.
>- You seem to be projecting it to ESPG: 2154, which is coordinate
>system applicable in France.( http://epsg.io/2154)  I can't see a
>Valid reason for doing this.
>- If you need to project to a projected Coordinate System,(Maybe for
>calculating areas & distances) you should use EPSG:7755, (
>http://epsg.io/7755) which is the one defined in NNRMS standards, and
>is used by many Government Organizations, like ISRO, NURM etc.
>- I generally* won't recommend* projecting to EPSG:3857, which is the
>webmercator projection, since it has several issues. *You should never
>measure areas or distances using this projection*.
>
>
>
>
> Regards,
> Devdatta
>
>
> On Tue, Apr 16, 2019 at 11:42 AM Nikhil VJ  wrote:
>
>> Hi Ashim,
>>
>> Pls provide the exact link of what you're referring - this stuff is never
>> as much on the top of people's minds as we assume it to be.
>> Do try loading your shapefiles on other tools and check there :
>> mapshaper.org is a site that does a lot of quick, cool things with
>> shapefiles; QGIS is a software that does everything with shapefiles. Both
>> are free and open source.
>>
>> The latter will help you re-save the shapefile into any CRS you want.
>> Right-click the layer > Save As (or "Export") and remember to choose your
>> preferred CRS in the dropdown.
>>
>> Can you give background on what this CRS : *2154* : is and why you want
>> to transform to that? Because I've only come across two main CRS's both
>> being under the "WSG 84" category :
>> - EPSG 4326 : this makes everything in latitude longitude
>> - EPSG 3857 : this makes everything in meters from the equator (I think)
>> and we need to get the data in this format when we want to do things in
>> physical distance terms like making a 1km distance buffer or measuring areas
>>
>> I haven't learned GIS stuff from theory, I just use it, so don't know
>> more details about CRS. I do understand that the dizzying multitude of
>> CRS's out there are so because apparently GIS folks like to re-orient the
>> center of the world (geometrically speaking) to where their data is to
>> ensure least distortion of their shapes in the rendering. When I come
>> across anything that's in a non-standard CRS, my first move is to transform
>> it to either lat-longs (4326) or meters (3857).
>>
>> Regards
>> Nikhil VJ
>>
>>
>> On Monday, April 15, 2019 at 1:57:20 PM UTC+5:30, Ashim wrote:
>>>
>>> Dear All,
>>>
>>> I am referring to 2011_Dist. * set of shapefile. (District level
>>> shapefiles).
>>>
>>> When I read them as simple features in R, ( like this )
>>>
>>> map <- st_read("2011_Dist.shp") %>% st_as_sf()
>>>
>>> ( it's unprojected because it says 4326 in the epsg when I read the
>>> above )
>>>
>>> plot(map$geometry)
>>>
>>> it looks OK.
>>>
>>> But when I do :
>>>
>>> map <- st_read("2011_Dist.shp") %>% st_as_sf() %>% st_transform(crs=
>>> 2154)
>>>
>>> plot(map$geometry)
>>>
>>> the map is TILTED.
>>>
>>> What is the correct projection to use for this dataset? Please clarify.
>>>
>>> Best Regards,
>>> Ashim
>>>
>> --
>> Datameet is a community of Data Science enthusiasts in India. Know more
>> about us by visiting http://datameet.org
>> ---
>> You received this message because you are subscribed to the Google Groups
>> "datameet" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to datameet+unsubscr...@googlegroups.com.
>> For more options, visit https://groups.google.com/d/optout.
>>
> --
> Datameet is a community of Data Science enthusiasts in India. Know more
> about us by visiting http://datameet.org
> ---
> You received this message because you are subscribed to the Google Groups
> "datameet" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to datameet+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

-- 
Datameet is a community of Data Science enthusiasts in India. Know more about 
us by visiting http://datameet.org
--- 
You received this message because you are subscribed to the Google Groups 
"datameet" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to datameet+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [datameet] Re: Correct CRS for datameet maps

[Ashim Kapoor]

Dear All,

I have a small query on similar lines. When do we use  UNPROJECTED maps ? I
understand that to go from 3D to 2D we need a projection. When is it
reasonable to use an unprojected map?

Many thanks,
Ashim

On Tue, Apr 16, 2019 at 12:12 PM Ashim Kapoor  wrote:

> Dear Devdatta and Nikhil,
>
> I attended a R GIS school where the teachers were from France and they
> were using CRS 2154. They told us that they were not familiar with the CRS
> for Indian maps.
>
> I used 2154 thinking that it would not make much  of a difference, but I
> was wrong. My apologies for the confusion. When I used that on Indian (
> datameet ) shapefiles,I got this : map1.png ( see attachment )
>
> I have also attached the R code to create these maps from the Datameet
> District level ( 2011 ) shapefiles. I have created map2.png and map3.png
> from crs = 3857 / 7755 respectively.
>
> I also found this :
> https://epsg.io/?q=india
>
> Which is confusing because many CRS are there for India. Why do we have so
> many choices of CRS for India ?
>
> I guess for the time being I will use 7755.
>
> Many thanks to Devdatta and Nikhil for their help.
> Best Regards,
> Ashim
>
>
>
> On Tue, Apr 16, 2019 at 11:42 AM Nikhil VJ  wrote:
>
>> Hi Ashim,
>>
>> Pls provide the exact link of what you're referring - this stuff is never
>> as much on the top of people's minds as we assume it to be.
>> Do try loading your shapefiles on other tools and check there :
>> mapshaper.org is a site that does a lot of quick, cool things with
>> shapefiles; QGIS is a software that does everything with shapefiles. Both
>> are free and open source.
>>
>> The latter will help you re-save the shapefile into any CRS you want.
>> Right-click the layer > Save As (or "Export") and remember to choose your
>> preferred CRS in the dropdown.
>>
>> Can you give background on what this CRS : *2154* : is and why you want
>> to transform to that? Because I've only come across two main CRS's both
>> being under the "WSG 84" category :
>> - EPSG 4326 : this makes everything in latitude longitude
>> - EPSG 3857 : this makes everything in meters from the equator (I think)
>> and we need to get the data in this format when we want to do things in
>> physical distance terms like making a 1km distance buffer or measuring areas
>>
>> I haven't learned GIS stuff from theory, I just use it, so don't know
>> more details about CRS. I do understand that the dizzying multitude of
>> CRS's out there are so because apparently GIS folks like to re-orient the
>> center of the world (geometrically speaking) to where their data is to
>> ensure least distortion of their shapes in the rendering. When I come
>> across anything that's in a non-standard CRS, my first move is to transform
>> it to either lat-longs (4326) or meters (3857).
>>
>> Regards
>> Nikhil VJ
>>
>>
>> On Monday, April 15, 2019 at 1:57:20 PM UTC+5:30, Ashim wrote:
>>>
>>> Dear All,
>>>
>>> I am referring to 2011_Dist. * set of shapefile. (District level
>>> shapefiles).
>>>
>>> When I read them as simple features in R, ( like this )
>>>
>>> map <- st_read("2011_Dist.shp") %>% st_as_sf()
>>>
>>> ( it's unprojected because it says 4326 in the epsg when I read the
>>> above )
>>>
>>> plot(map$geometry)
>>>
>>> it looks OK.
>>>
>>> But when I do :
>>>
>>> map <- st_read("2011_Dist.shp") %>% st_as_sf() %>% st_transform(crs=
>>> 2154)
>>>
>>> plot(map$geometry)
>>>
>>> the map is TILTED.
>>>
>>> What is the correct projection to use for this dataset? Please clarify.
>>>
>>> Best Regards,
>>> Ashim
>>>
>> --
>> Datameet is a community of Data Science enthusiasts in India. Know more
>> about us by visiting http://datameet.org
>> ---
>> You received this message because you are subscribed to the Google Groups
>> "datameet" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to datameet+unsubscr...@googlegroups.com.
>> For more options, visit https://groups.google.com/d/optout.
>>
>

-- 
Datameet is a community of Data Science enthusiasts in India. Know more about 
us by visiting http://datameet.org
--- 
You received this message because you are subscribed to the Google Groups 
"datameet" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to datameet+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [datameet] Re: Correct CRS for datameet maps

[Ashim Kapoor]

Dear Devdatta and Nikhil,

I attended a R GIS school where the teachers were from France and they were
using CRS 2154. They told us that they were not familiar with the CRS for
Indian maps.

I used 2154 thinking that it would not make much  of a difference, but I
was wrong. My apologies for the confusion. When I used that on Indian (
datameet ) shapefiles,I got this : map1.png ( see attachment )

I have also attached the R code to create these maps from the Datameet
District level ( 2011 ) shapefiles. I have created map2.png and map3.png
from crs = 3857 / 7755 respectively.

I also found this :
https://epsg.io/?q=india

Which is confusing because many CRS are there for India. Why do we have so
many choices of CRS for India ?

I guess for the time being I will use 7755.

Many thanks to Devdatta and Nikhil for their help.
Best Regards,
Ashim



On Tue, Apr 16, 2019 at 11:42 AM Nikhil VJ  wrote:

> Hi Ashim,
>
> Pls provide the exact link of what you're referring - this stuff is never
> as much on the top of people's minds as we assume it to be.
> Do try loading your shapefiles on other tools and check there :
> mapshaper.org is a site that does a lot of quick, cool things with
> shapefiles; QGIS is a software that does everything with shapefiles. Both
> are free and open source.
>
> The latter will help you re-save the shapefile into any CRS you want.
> Right-click the layer > Save As (or "Export") and remember to choose your
> preferred CRS in the dropdown.
>
> Can you give background on what this CRS : *2154* : is and why you want
> to transform to that? Because I've only come across two main CRS's both
> being under the "WSG 84" category :
> - EPSG 4326 : this makes everything in latitude longitude
> - EPSG 3857 : this makes everything in meters from the equator (I think)
> and we need to get the data in this format when we want to do things in
> physical distance terms like making a 1km distance buffer or measuring areas
>
> I haven't learned GIS stuff from theory, I just use it, so don't know more
> details about CRS. I do understand that the dizzying multitude of CRS's out
> there are so because apparently GIS folks like to re-orient the center of
> the world (geometrically speaking) to where their data is to ensure least
> distortion of their shapes in the rendering. When I come across anything
> that's in a non-standard CRS, my first move is to transform it to either
> lat-longs (4326) or meters (3857).
>
> Regards
> Nikhil VJ
>
>
> On Monday, April 15, 2019 at 1:57:20 PM UTC+5:30, Ashim wrote:
>>
>> Dear All,
>>
>> I am referring to 2011_Dist. * set of shapefile. (District level
>> shapefiles).
>>
>> When I read them as simple features in R, ( like this )
>>
>> map <- st_read("2011_Dist.shp") %>% st_as_sf()
>>
>> ( it's unprojected because it says 4326 in the epsg when I read the above
>> )
>>
>> plot(map$geometry)
>>
>> it looks OK.
>>
>> But when I do :
>>
>> map <- st_read("2011_Dist.shp") %>% st_as_sf() %>% st_transform(crs= 2154)
>>
>> plot(map$geometry)
>>
>> the map is TILTED.
>>
>> What is the correct projection to use for this dataset? Please clarify.
>>
>> Best Regards,
>> Ashim
>>
> --
> Datameet is a community of Data Science enthusiasts in India. Know more
> about us by visiting http://datameet.org
> ---
> You received this message because you are subscribed to the Google Groups
> "datameet" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to datameet+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

-- 
Datameet is a community of Data Science enthusiasts in India. Know more about 
us by visiting http://datameet.org
--- 
You received this message because you are subscribed to the Google Groups 
"datameet" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to datameet+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


map.R
Description: Binary data

Re: [datameet] Re: Correct CRS for datameet maps

[Devdatta Tengshe]

HI Ashim, & others,

Coordinate systems, and their area of applicability is an abstruse topic,
and one that has many complicated things to consider. So what I'm saying
here is a user-friendly summary.


   - The data in our Repository is in WGS-84 Lat/long, which has the EPSG
   code 4326.
   - You seem to be projecting it to ESPG: 2154, which is coordinate system
   applicable in France.( http://epsg.io/2154)  I can't see a Valid reason
   for doing this.
   - If you need to project to a projected Coordinate System,(Maybe for
   calculating areas & distances) you should use EPSG:7755, (
   http://epsg.io/7755) which is the one defined in NNRMS standards, and is
   used by many Government Organizations, like ISRO, NURM etc.
   - I generally* won't recommend* projecting to EPSG:3857, which is the
   webmercator projection, since it has several issues. *You should never
   measure areas or distances using this projection*.




Regards,
Devdatta


On Tue, Apr 16, 2019 at 11:42 AM Nikhil VJ  wrote:

> Hi Ashim,
>
> Pls provide the exact link of what you're referring - this stuff is never
> as much on the top of people's minds as we assume it to be.
> Do try loading your shapefiles on other tools and check there :
> mapshaper.org is a site that does a lot of quick, cool things with
> shapefiles; QGIS is a software that does everything with shapefiles. Both
> are free and open source.
>
> The latter will help you re-save the shapefile into any CRS you want.
> Right-click the layer > Save As (or "Export") and remember to choose your
> preferred CRS in the dropdown.
>
> Can you give background on what this CRS : *2154* : is and why you want
> to transform to that? Because I've only come across two main CRS's both
> being under the "WSG 84" category :
> - EPSG 4326 : this makes everything in latitude longitude
> - EPSG 3857 : this makes everything in meters from the equator (I think)
> and we need to get the data in this format when we want to do things in
> physical distance terms like making a 1km distance buffer or measuring areas
>
> I haven't learned GIS stuff from theory, I just use it, so don't know more
> details about CRS. I do understand that the dizzying multitude of CRS's out
> there are so because apparently GIS folks like to re-orient the center of
> the world (geometrically speaking) to where their data is to ensure least
> distortion of their shapes in the rendering. When I come across anything
> that's in a non-standard CRS, my first move is to transform it to either
> lat-longs (4326) or meters (3857).
>
> Regards
> Nikhil VJ
>
>
> On Monday, April 15, 2019 at 1:57:20 PM UTC+5:30, Ashim wrote:
>>
>> Dear All,
>>
>> I am referring to 2011_Dist. * set of shapefile. (District level
>> shapefiles).
>>
>> When I read them as simple features in R, ( like this )
>>
>> map <- st_read("2011_Dist.shp") %>% st_as_sf()
>>
>> ( it's unprojected because it says 4326 in the epsg when I read the above
>> )
>>
>> plot(map$geometry)
>>
>> it looks OK.
>>
>> But when I do :
>>
>> map <- st_read("2011_Dist.shp") %>% st_as_sf() %>% st_transform(crs= 2154)
>>
>> plot(map$geometry)
>>
>> the map is TILTED.
>>
>> What is the correct projection to use for this dataset? Please clarify.
>>
>> Best Regards,
>> Ashim
>>
> --
> Datameet is a community of Data Science enthusiasts in India. Know more
> about us by visiting http://datameet.org
> ---
> You received this message because you are subscribed to the Google Groups
> "datameet" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to datameet+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

-- 
Datameet is a community of Data Science enthusiasts in India. Know more about 
us by visiting http://datameet.org
--- 
You received this message because you are subscribed to the Google Groups 
"datameet" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to datameet+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.