My team is working with demographics data across different data sources (some paid sources and some free data sources available online).

Each of these data sources comes with a shapefile and some attributes associated with each demographic area and could be defined across different cuts of time.

However, when we display these attributes to our end users, we wanted to abstract the multiple datasources concept and show zip codes as a single demographic unit.

We were planning to combine the attributes of all the datasources into one single data source and point that to one of the shapefiles (For the time being, we are willing to look past issues related to granularity or precision in the definition of polygon across these data sources.

However, should we be concerned that the ZIP code to actual geographic area might not be consistent across demographics datasources taken across different cuts in time? E.g. Zip code 12345 used to Map to an area in State A till 2010 but points to an area in State B for all datasets after 2010?

closed as primarily opinion-based by John Powell, user30184, Mapperz Mar 1 '16 at 17:03

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    Yes, you should be concerned that you would be misrepresenting data. Are you wanting to show the most recent data for any given location? If so, let us know what software you're using and we can give you some ideas of how to address the problem. – Tom Feb 29 '16 at 20:59
  • 1
    You shouldn't be so concerned with the zip code itself. Remember that in spatial analysis, location often functions as your primary key. That is, if you have the shapefiles associated with each set of tabular data, then you don't need to sacrifice anything. Instead, simply union all your shapefiles, along with their attributes. You might want to change field names first, so that--for example--the date of an attribute isn't indecipherable to your end users. – Tom Feb 29 '16 at 21:16
  • Zip codes are tricky. Like you said they will change over time, in fact I believe they change rather often (sometimes on a monthly basis) and because of this your results may not be accurate. Are there other unique fields in this data other than zip code? What time periods are you looking at? Also can we maybe see a slice of the data you are talking about. – ed.hank Mar 1 '16 at 0:25

I agree that it should be an issue of concern, especially if you're normalizing your data using geography. There are several different techniques for disggregating data using area weighting models. In addition, if you're actually using Zip Code Tabulation Areas (ZCTA) and the other data are population based, you could use zip code crosswalk tables to distribute your estimates to a more granular geography like Tracts.

Not the answer you're looking for? Browse other questions tagged or ask your own question.