topic Re: Combining data with different scales in Data Management Questions

Combining data with different scales

anayoung — Fri, 18 Dec 2015 15:04:27 GMT

Hello,

I have two feature classes that I want to combine into one feature class but they were created in two different scales. My thought was that I simply loaded both into one feature class and call it done; however, I realize there might be a better way to get this done. Any suggestions?

Thanks,

Ana

Re: Combining data with different scales

DanPatterson_Retired — Fri, 18 Dec 2015 15:21:42 GMT

why not leave them separate? Is there a reason that they have to be combined?

Re: Combining data with different scales

WesMiller — Fri, 18 Dec 2015 15:22:23 GMT

What do you mean when you say different scales does that mean different projections?

Re: Combining data with different scales

anayoung — Fri, 18 Dec 2015 15:32:48 GMT

Both feature classes represent polygons of city boundaries. One feature class is from the Census data and the other is polygon data I gathered directly from the cities themselves. The idea is to delete features from the Census data and replace them with the data I received.

The data is representing the same thing (i.e. cities) so it would be good to have them combined.

Re: Combining data with different scales

anayoung — Fri, 18 Dec 2015 15:36:35 GMT

Both feature classes are in the same projected coordinate system. However, they were created at different scales. One set is from the Census which was created at a standard 1:100,000 scale and the other is a feature class I created using data I received from multiple sources which looks fine. But I am concerned about combining the Census data with the feature class I created. Is there anything I should be aware of or do when I combine them?

Re: Combining data with different scales

DanPatterson_Retired — Fri, 18 Dec 2015 15:38:43 GMT

if the geometries are identical, just to a spatial or attribute join to bring the attributes over.

If the geometries are different, then be aware that you could end up with a whole load of boundary issues.

It would be better to leave the original files alone and combine, if necessary, the good bits of both files, you then have the originals and it is easier to put something together than to take something apart

Re: Combining data with different scales

ChrisDonohue__GISP — Fri, 18 Dec 2015 15:41:20 GMT

One question that arises is why? If you are using Census data, why would you want to essentially "move" it to the new city spatial data? Where do you want to go with this? I ask this not as a criticism, but as the reason why could very much affect the outcome. For example, Census data is created with certain assumptions and constraints (including spatial). I'm not an expert on Census data, but I suspect changing the spatial part may invalidate some of it, so it would be best to know where this is going so as to be able to evaluate whether the outcome will still be valid.

Chris Donohue, GISP

Re: Combining data with different scales

WesMiller — Fri, 18 Dec 2015 15:54:53 GMT

If you have an Advanced license you may be interested in the Symmetrical Difference—Help | ArcGIS for Desktop sounds like it's BAS time

Re: Combining data with different scales

anayoung — Fri, 18 Dec 2015 15:57:42 GMT

A spatial join would indeed create a lot of boundary issues. So, I'll just plan on combining the good bits together and, of course, keep the originals separate. Thanks!

Re: Combining data with different scales

anayoung — Fri, 18 Dec 2015 16:05:52 GMT

Yes, symmetrical difference would be helpful if I planned on keeping the Census boundaries of the cities I collected. The Census data has over 300 cites but my data only has 40 cities (better quality boundaries). My plan is to remove the 40 cities from the Census data and replace them with my 40. So I'll at least have 40 cities with better boundaries. Does this make sense?

Re: Combining data with different scales

DanPatterson_Retired — Fri, 18 Dec 2015 16:14:46 GMT

Just be aware of the boundaries. I would symbolize one of the layers with just an outline and have a look at the edges. You don't want to cut out part of the old layer if the new layer has changed borders. And symmetrical difference isn't what you want, it would be a union which would keep all the good bits and fiddly bits and enable you to decide what to keep and what to discard. Also check intersect and other tools in the Overlays tools An overview of the Overlay toolset—Help | ArcGIS for Desktop

Re: Combining data with different scales

anayoung — Fri, 18 Dec 2015 16:26:54 GMT

Thank you, Dan. Good advice. I will do as you suggest.

Re: Combining data with different scales

ChadKopplin — Fri, 18 Dec 2015 17:07:16 GMT

Also, keep in mind that you are setting your scale to the broader of the 2 datasets, when you combine them. This could affect the accuracy of your final product, and any other analysis that you may use the product in.

Re: Combining data with different scales

anayoung — Fri, 18 Dec 2015 17:40:15 GMT

That is my greater concern. I apologize I wasn't able to communicate it properly initially. Is there a way to cut down on any possible issues?

Re: Combining data with different scales

katherinehannaford — Mon, 28 Dec 2015 16:29:04 GMT

As an addendum to this interesting discussion:

If your cities do not share boundaries, then your boundary issues shouldn't be a problem in a combined dataset. If they do share boundaries, then, as previously stated, you will have to be careful or decide if you are going to change the adjacent Census city boundary. And, if you ever spatially nest Census tracts or blocks onto your cities (such as on a map) then the boundary differences will also be problematic.

If you do combine the census and local sources I would recommend populating a new attribute to keep a record of the source (census2010, CityA, CountyX, etc). How you actually combine the datasets will depend a lot on what attributes you want to keep from each set, so you'll want to keep an eye on that. And, whatever you do, make sure you're editing in a new dataset or a copy of an existing one, just to be on the safe side.

Good luck.

Re: Combining data with different scales

anayoung — Mon, 28 Dec 2015 21:59:55 GMT

Hello Katherine,

Thank you for your thoughtful response.

I made sure to get local data for those cities that share boundaries from one source e.g. Regional planning commissions. So, that shouldn't be a problem. I also created a new feature class with a source (of data) field. However, I am concerned about the differences in scale between the Census data and local data as the Census is created at a much smaller scale. Thank you for your insight on creating a map with this combined data and census tracts and blocks.

Ana

Sent from my iPhone

Re: Combining data with different scales

katherinehannaford — Tue, 29 Dec 2015 15:23:25 GMT

Ana,

The scale differences per se aren’t a problem – the scale of your combined dataset is the smaller scale of the census data. However, interpretation problems could come up if the boundary description is different.

As an aside, I will mention that Census does try to elicit boundary files from local government, such as county or regional planning organizations in the run-up to the decennial census (PSAP). Theoretically, they should be the same, although I know from personal experience that this is not always the case.

Good luck with your work,

Kathi

Kathi Hannaford

SEDA-COG GIS

570-522-7288

khannaford@seda-cog.org<mailto:khannaford@seda-cog.org>

“Start where you are. Use what you have. Do what you can.”

Arthur Ashe

Re: Combining data with different scales

anayoung — Tue, 29 Dec 2015 16:25:49 GMT

Thank you, Kathi. I've combined data sets with different scales before and had no problems. However, I wasn't considering the fact that the combined set might be used in advanced analyses. So, I needed some confirmation that the data would be okay should it be used for more than just display and identification of attributes. I think you have put my worries to rest.

With much appreciation,

Ana

Sent from my iPhone