For your csv/Excel data, you can use the Add XY Data tool: File menu, Add Data, Add XY Data. After you browse to the file and select it, the tool will default the coordinate system to that of the maps. You probably instead want to override that and choose Geographic Coordinate System, North America, NAD 1983 (or North American 1983 depending on ArcGIS version).
Once the XY layer is created, you can right-click the layer and choose data, export data. In that tool, you can write the data out to a feature class, or a shapefile. There's an option to use the data's coordinate system (NAD 1983) or use the data frame's coordinate system.
To set the data frame's coordinate system to NAD 1983 UTM 18 North, choose projected coordinate system, UTM, NAD 1983. Or in the search box, type in 26918, which should select it for you.
Unless you are covering a large area, why are you using that coordinate system? In any event, you should project the data using an appropriate projection for your area. For 6 degree latitudinal widths, a UTM fits most bills, or Lambert Conformal Conic for larger EW extents.
I guess it depends on what you mean by large area, state of NY, but I wanted this because the map looks better, in my opinion, as it does not make my shape file look flattened.
I'm new to this, so you will have to explain what you mean as I don't understand.
Carter, this link provided by New York State cites several data sources as well as the coordinate system/projection that the data are available in. In a pinch use UTM Zone 18 (Data Management Tools, Projections and Transformations, Project tool NOT the define projection tool) and make it so since NY is 'mostly' in UTM zone 18. There are other state plane systems for NY, which are more accurate but their geographic is limited to a smaller area. So have a look at the link
Not quite clear on ‘to interpolate data from latitude and longitude points’.
Nevertheless, I assume that you want to spatially interpolate data from sparse points in geodetic coordinates (Long, Lat), for example, air pollution or soil moistures, etc.
If so, your measurements’ interpolation by Geostatistics technique can be done spatially in any projected or geodetic coordinate systems like local, UTM or WGS 1984 Web Mercator Auxiliary Sphere. To decide which coordinate
system to use, it fully relies on the area and project you are working on (affecting your mapping accuracy only).
Secondly, if you want to make your interpolation successful, you are mostly required to ‘properly’ model spatial relationship among your measurements first, before taking interpolation, especially, with variant Kriging.
Thirdly, conduct the uncertainty analysis on your interpolation...
Have a look at Empirical Bayesian Kriging, which in 10,3, makes use of chordal distances if your data are in an unprojected coordinate system.
Or do your interpolation in an appropriate projected coordinate system and project the output to GCS, however, often the area is too large for this and that is why we introduced chordal distances.
Ok let me explain this a bitter better. I can get it to plot the XY coordinates correctly (Image 1) but when I use the wizard to interpolate the data, the project ends up correct but in some other place really far away. (Image 2)
From your plot, it looks that your sparse data already are well distributed in the map. However, you are not satisified with your interpolation results (by the Wizard). Correct?
If so, as said, nice to put more effort on modeling spatial relation among your data (rather than coordinate system), before using the Wizard.
Honestly, on our expereince, the GA Wizard in ArcGIS offers an easy workflow for some of users. But, mostly it is not case, if only using parameters by default.
If your data are still not mapped into the map 'properly', because you don't know the exact coordinate system of coodinates in the data from the CSV file, you can attach and share here.