For spatial analysis purposes, do "points" need to also be in a common coordinate system?

[Craig_Eissler_Iceman]

I understand the best practice for having operational layers, and the Map itself, in the same projected coordinate system for spatial analysis purposes. Since 'lines and polygons' have length and/or area, I can see why this could matter.

However, what about 'points' ...I'm thinking they would be exempt from a potential scaling issue due to being in a different coordinate system, but maybe I'm wrong..? Or, maybe they just need to be in the same datum?

Thanks, CCE

[DanPatterson]

same units (degrees, meters, feet) and same datum

if in degrees, you have the option of geodesic perimeter and area for poly* features (where appropriate) and proximity/distance for points and poly* features

if in planar units, then you have planar versions of the above (i.e. euclidean space)

projections of both geographic and planar data permit provide all combinations.

points, lines or polygons don't suffer from scaling issues, coordinate systems do. a projectected coordinate system can be chosen to preserve either: direction, distance, area or an *ish combination of the aforementioned (eg. a conformal coordinate system like UTM).

For small areas, I use a conformal system with a scale factor of 0.9999 (MTM, state plane) since it is a compromise.  If I need one of the map properties preserved, then I project the data to an appropriate coordinate system, or back to geographic and use geodesic calculations.

If I want a pretty map, I choose a projection that suits the point I am trying to make

[Craig_Eissler_Iceman]

Thanks, but I was hoping we could start by answering as a "Yes or No"
question?

[DanPatterson]

..I'm thinking they would be exempt from a potential scaling issue due to being in a different coordinate system, but maybe I'm wrong..?  yes, you are wrong

Or, maybe they just need to be in the same datum  yes, you are partially right

[Craig_Eissler_Iceman]

Thanks. ...OK, so points ALSO need to be projected to the same coordinate system as the other operational layers to for accurate spatial analysis.

Just to confirm, is this true?
If not, please correct my statement.

[DanPatterson · Accepted solution]

All geometry, vector and raster, should be in the same coordinate system.  It is the unwise person that relies on "projection on the fly" to do any analysis, rely on "human intelligence" and overt decisions about what you embark on

[Craig_Eissler_Iceman]

Thanks, understood; I was only unsure about "points" since they have no dimension per se', but OK I see what you're saying that they would still be projected-on-the-fly, and therein lies the potential problem. 

So, I'll re-project points as well!