I have downloaded a couple of data sets and am noticing that the "sectors" fields for these data sets do not always correlate exactly to the IATI terminology in the Sector Code List:  http://iatistandard.org/codelists/sector 

Example:  Environmental education/ training; Flood protection; Public administration- Water, sanitation and flood protection; Natural disaster management

In the example above, "Environmental education/ training" is found in the code list, but "Flood protection" is not.  

Am I missing something?  

Hello Brandon,

The IATI Standard allows donors to publish with multiple types of sector classification. 

This should be indicated by a 'vocabulary' attribute on the <sector> element in the XML. Only things with @vocabulary='DAC' have to comply with the code list as I understand.

This allows flexibility for donors who do not yet store their data against the DAC codelist, but of course, it can make comparing across files more complex. There are some plans in the pipeline at AidInfo for a project to help with mapping between different codelists...

If you are finding donors publishing non DAC codes when they declare the vocabulary is DAC, then it's worth reporting that as an error with the particular donors file. You should be able to do that here on the KB and then the quality control team working with donors will be alerted to look into it...

Thanks Tim --

"The IATI Standard allows donors to publish with multiple types of sector classification."

If that's the case, then what is the purpose of the IATI specific sector codes?  

FYI, the data set I am referencing shows the following (in the .csv file):   sector_vocabularies = DAC; WBSector; WBTheme.  

Hey Brandon,

I believe the goal is that over time donors are encouraged to converge on the IATI Sector Classifications (and those classifications themselves are likely to be refined over time), but in the initial development, a lot of what is published will be based on how donors existing systems have data classified. So donors who already report to the OECD's Development Assistance Committee (DAC) have those classifications assigned to their data, but others don't yet.

In the World Bank file you're referencing above, what you in fact have is two classification schemes in operation. World Bank files originally were only coded with their internal sectors and themes, but they have added a mapping to DAC classifications too. 

The way the CSV serialisation works, the sector_vocabularies column should a ; separated list of vocabularies in the same order as the sector codes. So, if you were only interested in DAC codes, you can ignore any sector codes/names in the same position in their respective columns as WBSector and WBTheme. 

If this proves difficult to do for the sort of analysis you are trying, feel free to describe in more depth your use case for the data and I'll be happy to see if there might be a good way to get the data into the shape you want...

Tim -- Very helpful, thank you.  A couple of follow-on questions then:

• Are there any data sets within the IATI Registry whose value for sector_vocabularies is "IATI"?
• In the IATI Sector Code list, there is an element called Category (numeric) and an element called Category Name (text).  What do these refer to?  They are separate from the Name and Code elements, which appear to refer to the sector name and sector code.

Hello Brandon,

Sorry for the slow reply: been away. 

Where people are using the codelist from the IATI Standard site they should be using 'DAC' as the vocabulary name. There isn't a sector vocabulary called 'IATI' at present. I'll mention to the documentation team that they might need to document this more clearly.

In the IATI Sector Code List, the Category and Category name elements reflect that the DAC codelist is, to some extents, hierarchical, so that sectors can be grouped together in their categories.