| Geographical jittering over time [message #28170] |
Tue, 21 November 2023 10:28  |
pedrorosado
Messages: 2
Registered: October 2023
|
Member |
|
|
Hello all,
I'm using multiple surveys, of different countries and different years, as well as their Geographical data, to create a map that covers a good portion of Africa. Now, I know that regarding geographical data, DHS uses some masking, particularly a random jittering. Reading some documents, particularly one from 2013, I find that urban areas are subject to a 2km, while rural are subjected to a 5 km jitter. My question is: is this constant over time? Were surveys pre-2013 also subject to this rule, as well as every other survey post-2013?
Now, I know that post phase VII (post-2013), for each .zip of Geographical data there is a readme with this information, but given the very large number of surveys that I'm currently using, it would not be efficient if I were to open every single one of them, as well as code every single exception. As such, I would be very thankful if anyone had a good answer to this.
Br,
Pedro
|
|
|
|
| Re: Geographical jittering over time [message #28235 is a reply to message #28170] |
Mon, 04 December 2023 13:04  |
Rose-DHS
Messages: 8
Registered: October 2021
|
Member |
|
|
Georeferenced surveys from 2003 onwards are displaced using the standard displacement procedure described in SAR7 (https://dhsprogram.com/pubs/pdf/SAR7/SAR7.pdf). Urban clusters are displaced up to 2km, while rural clusters are displaced up to 5km with a further 1% of rural clusters displaced up to 10km.
Surveys conducted prior to 2003 were not displaced using the standard displacement procedure. Coordinates for the earliest surveys were obtained from paper maps, gazetteers of settlement names, or preexisting census data files, while GPS collection began in 1996. The method used to determine the lat/long coordinates for each cluster is listed under the SOURCE attribute in the GE datasets.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
