dhis2-devs-core team mailing list archive

[Lars Helge Øverland <larshelge@xxxxxxxxx>]

Hi,

quick summary from California trip. Jan and I went on a 3 day "expert
meeting" at google. First part was training on the earth engine platform
<https://earthengine.google.org/#intro>. This is a cloud computational
platform that makes available a ton of satellite images e.g. from the
landsat <http://landsat.gsfc.nasa.gov/> program and allows you to do
analysis like reducing, aggregating, filtering and combining multiple
images into layers on google data centers. Lots of researches from US and
European agencies and universities that made presentations on how they use
it.

Second part was a discussions around how this engine could be more useful
for researches and "tool makers". Google has committed to using 1% of their
resources for "good". So far the earth engine has been used mostly for
forestry / deforestation. They now want to apply it within health. The most
typical use-case seems to be mapping of risk of malaria and other epidemic
prone diseases. We met engineers from the Google maps, earth, earth engine
teams who wanted feedback on which features would be useful in the health
domain. We also got to present - it was cool to present DHIS 2 in the
middle of Googleplex ;)

In our group there was people from health metrics and evaluation
<http://www.healthdata.org/>, UCSF <https://www.ucsf.edu/>, Gates found,
CHAI, Oxford uni, RTI, Carter center
<http://www.cartercenter.org/index.html>, Uni of Pittsburgh
<http://www.pitt.edu/>. There was actually lots of talk of DHIS 2, many
people knew about it and it seems Google might develop a DHIS 2-compatible
import tool for the earth engine.

Last day we met with Hugh Sturrock, Adam Bennet, Roly Gosling at UCSF (very
nice people). We will now work together on an app for mapping of malaria
risk, probably in zim. The idea is interesting - it will detect malaria
risk based on rainfall, elevation, vegetation, moisture (this data can be
derived from satellite images or come from other data sets in the earth
engine), then overlay it with household visit data collected using the new
DHIS 2 tracker android app (data on house spraying, bednets, prophylaxis).
This should tell you exactly which areas are at risk, which houses have not
yet been sprayed in last x months, and hence where you need to go next.
UCSF will hire vizzuality <http://www.vizzuality.com/> to do the app
coding. This means we need to feed data into earth engine so google will
support on that end. Jan and I have discussed to integrate earth engine
satellite images directly as a layer in GIS as well.

There is still an "if" in that UCSF must get their funding request approved
by gates, but otherwise this looks pretty promising.

Lars