Research DHS, MICS APIs for Gisida Ingestion

[cKellyDesign]

As a Gisida client I should be able to connect layers data sources to DHIS and MICS data.

This ticket tracks research outcomes, insights, and conversations.

[cKellyDesign]

DHS API

• The API itself is public (I think) so Gisida could call it directly by specifying an API URL as the layer data source
• The API has good documentation on all of it's endpoints and can be accessed as REST or Query based formatted URLs (http://api.dhsprogram.com/rest/dhs/[ENDPOINT]/[PARAM_VALUE] is the same as http://api.dhsprogram.com/rest/dhs/[ENDPOINT]?[PARAM_KEY]=[PARAM_VALUE])
• The API is flexible and can be utilized in a way to dynamically load DHS data (http://api.dhsprogram.com/rest/dhs/[ENDPOINT]/[PARAM_1],[PARAM_2],[PARAM_N])
• API response i paginated (100 records per page) but if we pass in our API key the we can get responses with more reports per page, otherwise we can deal with pagination
• API can return JSON (by default) or CSV formatted data, JSON will probably be easier

I do not yet understand how to query subnational level data.

Endpoints

Countries - http://api.dhsprogram.com/rest/dhs/countries Indicators - http://api.dhsprogram.com/rest/dhs/indicators Surveys - http://api.dhsprogram.com/rest/dhs/surveys Data - http://api.dhsprogram.com/rest/dhs/data

Next Steps

1) Immediate - Attempt a country level layer in Service Mapping (will require a country level geojson file or shapefile/vector tileset to join with for fill layers) 2) Short Term - If 1 doesn't work, use the API endpoints to build out static CSV/JSON files for Gisida layers to be loaded in normally 3) Long Term - Build a DHS connector module in Gisida to ingest API responses for layers

[cKellyDesign]

MICS

I cannot seem to find an API for this data nor can I find any data output or downloads other than for statistical analysis.

Survey List - http://mics.unicef.org/surveys MICS Data Tools - http://mics.unicef.org/tools?round=mics5#data-processing

@csc5k @engwiri - Do we need to produce CSV or JSON outputs from the datasets we've been given access to?? I hope not because that sounds like a lot of manual work and requires knowing SAS/STAT

This doesn't seem like it would be live data.

[csc5k]

@cKellyDesign DHS - I found a "RegionId" from the first API listed here: http://api.dhsprogram.com/#/introapi.cfm

I also see a "RegionId" field from the query-based API output: http://api.dhsprogram.com/rest/dhs/data?countryIds=KE&f=xml

I think that might give us the subnational breakdowns

Regarding MICS - I'm not sure what you mean about producing CSV or JSON outputs. I don't think we need to have the Gisida site produce that FOR the client, but we do need to translate SAS/STAT formats into CSV so that Gisida can use it. Is that what you mean?

[engwiri]

@cKellyDesign i think what we need immediately is to understand feasibility and possible level of effort. It is fantastic that DHS has an API thats a great first step. Do you know if you can access shape files? or would that have to be separate either way i don't think thats a big deal. We do need to have a sense of how much it would cost to build the long term solution for country and sub national levels

[cKellyDesign]

@csc5k - Yeah I saw the RegionId but I think that is actually multiple countries per region (at least that was my understanding without looking into it further). As for MICS, correct Gisida needs the CSV to work.

@engwiri - I didn't try pulling the geographic data along with the indicator data since I figured we could just get the shapefiles from HDX since we will need to combine them anyways.

As for LOE of a DHIS connector module in Gisida I estimate about 5-8 days of engineering work.