A picture of what to do...

I apologize for the cliche, but I want to draw out a picture with my words...

I see the efforts of the NCPHI Grid lab progressing towards sort of a city map of options and interests, and being that I live in Atlanta, it's (naturally) playing out as a wagon wheel with a big network of semi-clogged roads surrounded by a ring bypass with a light rail system that is trying it's best.

The way I see it, the suburbs and exurbs are like the user interfaces, and the downtown areas are the land of data.

The exurbs are like the user interfaces for people who don't really need all the data, just a summary... It would represent things like the "health channel" or widgets on the front page of "health.gov": Things that turn red if something bad is happening... or things that will send out an email if you need to be extra careful about washing your hands at a particular football stadium. The inner-city suburbs are more like maintenance interfaces where really hard-core statisticians go to conduct some hard-core informatics and data mining.

Right now, my impression of this city is that there are two or three really big skyscrapers (like Biosense databases) and a bunch of little warehouses (hospitals). But the problem is that all the roads between these things are gravel and the parking at the big skyscrapers is expensive. In the meantime all the secure trucks keep sinking and get stuck on rainy days, and no-one has really build up sets of nice subdivisions anywhere.

In the meantime, NCPHI is sort of building model homes in certain spots, paving over a few of the roads to make reach out on a few points (AMDS-GMAP/Rodsadai and NPDS-GMAPS/poicondai being suburbs on the north end) and building a train station at either end...

I feel like the end point we are trying to get to is filling in the map.

On the outside, lots of new UI's need to be developed, some really complex ones with lots of analysts, some very simple ones so that people can essentially Google their neighborhood to see if anything is amiss, lots of support to make such things automatically set up and install instead of being a 17 step rollout.

In the city, a lot of virtual and physical infrastructure needs to be set into place: connecting all the data silos and little warehouses together. Even having little services run around indexing and enriching (instead of this data point only having a zip code and an address, it now has a zip code, address, county, census track...) and backing up the data (so now if one service goes down it can be modified by other services), and also lots of different pathways to share the data, with large pipes for big file transfers and secure networks for identified patient data, and all the connections to get the data out to the UI suburbs.

That is sort of where I see thing should be, and I also don't think NCPHI can really build all of this, but can really just plot out the map and tell people "it would be really cool if you turned this field of nothing into a really cool alert application using this tech connecting to these services..." or going "make this but for this style of data coming from this aggregator".

So yeah, a little thought on how all this is set up.