Fwd: converging work streams for all SeaNET features

[John Tigue]

Begin forwarded message:

From: John Tigue <jo...@tigue.com>

Subject: Re: converging work streams for all SeaNET features

Date: May 21, 2015 9:20:19 AM PDT

To: Chris Ritzo <cri...@opentechinstitute.org>

Cc: Nathan Kinkade <kin...@opentechinstitute.org>, Collin Anderson <col...@gmail.com>, Georgia Bullen <geo...@opentechinstitute.org>, Bruce Blood <Bruce...@seattle.gov>

Looping in Bruce. My comment all collected at the bottom of this message.

On May 21, 2015, at 7:23 AM, Chris Ritzo <cri...@opentechinstitute.org> wrote:

Good morning Nathan, John & Collin:

Following our meeting with Bruce and team at the city earlier this week,
I wanted to send a quick note to coordinate and focus tasks for
completing the work on the Seattle Map and test. Collin for your
reference, the SEAnet work has been taking place here:
https://github.com/codeforseattle/seanetmap

and requirements are here:
https://github.com/codeforseattle/seanetmap/wiki/Requirements-from-the-City-of-Seattle,-Annotated

As I mentioned on the call, we have had a contractor (David) working on
a turnkey solution that, combined with Nathan and John's work on the
current geojson based map, and Collin's assistance on a customized test,
we should be able to get to the full set of requirements for the city
very easily.

David's deliverable is a deployable VM image, that:
- includes setup scripts to prepare a postgres database and python webserver
- uses configuration values from a text file such as: location (ie:
Seattle), google account to be used with the Big Query API, etc.
- once configured will pull in data from M-Lab via the Big Query API and
populate the postgres database
- provide a couple examples of visualizing that data via postgres

Nathan has been in separate meetings with David, and I'll send a
separate introduction email to bring David into this conversation, but
here's what I'd like to see happen next:

- Chris will discuss with David sharing his code in progress with this
group, specifically the postgres schema
- David/Nathan/John/Chris will discuss where to deploy a live test
instance of the VM
- Nathan & John work on making the bq2geojson map be driven from David's
postgres database
- Collin will customize the NDT websockets test that will both submit
the normal NDT results to M-Lab AND collect additional values in a table
in the local postgres db:
-- NDT test ID (to compare M-Lab BQ data with additional/more specific
local data)
-- HTML5 geolocation
-- Location type (e.g. residence, business, public space)
-- Connection type (Cable, DSL, Fiber, Cellular, Other, Unknown)
-- Advertised speeds, if known
-- Cost of service, if known
- The final map should display M-Lab data and where applicable the
locally collected data above. I'm imagining that the map will look at
M-Lab data from Big Query and IF matching NDT test IDs are found in
postgres, will use that more specific lat/long data and mash in the
additional locally collected fields in aggregate

We'll discuss more on our weekly call next week, but I wanted to get
this out to provide my vision and direction for merging these work
streams with the goal of providing this VM to Seattle for their
broadband project.

If you have any questions, please let me know.

-Chris

--

Chris Ritzo
Interim Director of Technology
Project Manager, M-LAB http://measurementlab.net
Open Technology Institute @ New America
1899 L St. NW Suite 400   | 202-596-3406
Washington, DC 20036
Public Key: 0x085B309C
8859 E047 852C C3E4 F421 04DD 0285 4037 085B 309C

That all sounds good and it looks like there is tech (David's work) which fills the big gaping hole of "no database in SEANetMap, and if there were it would need to sync with M-Lab".

I have a document I was pretty far along on that maps out various options to solve that but with this plan it is moot so I will stop working on it.

Tigue refocusing on Angular for front-end

Given what was described, it sounds like the simplest thing in terms of coordination would be for me to focus my energies on the front end until David's and Collin's work is integrated. My idea there has been to strap on an Angular context for Nathan's work. This would also involve Google's new Material Design implementation, angular-material:

https://github.com/angular/material

The above is simply Material Design implement atop Angular. It will give SEANetMap that fresh fashionable design aesthetic which actually is pretty nice. It standardized as much as possible the UX across Android and web apps. Google is trying to address the perceived quality gap between their offerings and Apple's. That is a tall order but Material Design really does help, not completely but a lot. More to the point here: tons of free code upon which to build.

The seed codebase is called material-start:

https://github.com/angular/material-start

That repo's readme has an image which I will reference here to explain what I have been thinking about.

What will the SEANetMap front end look like? 

The main pane as shown in the material-start repo's readme would contain Nathan's Leaflet map. 

In material-start's readme you can see a sidebar of available items, which retreats behind a hamburger stack on small mobile screens. In the material-start demo the items are people. In SEANetMap they will be (with better labels of course):

1. Run an NDT test (CityReq 1)

2. Enter/viiew details about your provider (CityReq 1.2)

3. Multiple buttons to city-wide report charts per provider (Average speed/latency, Performance by time of day, Wired vs. wireless performance (CityReq3.2)

The various dialogs (test results with authorization to submit them and the various per-provider charts) will be "pop-up" as shown in the material-start repo's readme. I see these as being D3 implement but that is another discussion. This plan just addresses the front-end big picture.

The primary coordination point in this plan is with Colin's customized NDT test. I will work on a feature branch which will bring in the basis of what I just outlined. I hope to have that far enough along by Tuesday such that Colin would be able to wire in his work, and so that we can simply see it for Tuesday's meeting and decide if we want to move forward with this plan.

-John

  

[John Tigue]

Thank you, Bruce.

I will keep being the point person on this project from the Code for Seattle side. 

This is cool M-Lab tech that we are deploying locally.

I am really excited to be involved in an open source project which is looking like it will impact the real-world.

-John

Begin forwarded message:

From: "Blood, Bruce" <Bruce...@seattle.gov>

Subject: RE: converging work streams for all SeaNET features

Date: May 21, 2015 12:04:40 PM PDT

To: John Tigue <jo...@tigue.com>, Chris Ritzo <cri...@opentechinstitute.org>

Cc: Nathan Kinkade <kin...@opentechinstitute.org>, Collin Anderson <col...@gmail.com>, Georgia Bullen <geo...@opentechinstitute.org>

Thanks, John.

 

A quick note on roles on the Seattle side, folks:

 

Bruce – City of Seattle project sponsor and business lead

John – Project development lead

 

Thanks, all!  This definitely feels like progress!

 

-BB

 

From: John Tigue [mailto:john@tigue.com] 
Sent: Thursday, May 21, 2015 9:20 AM
To: Chris Ritzo
Cc: Nathan Kinkade; Collin Anderson; Georgia Bullen; Blood, Bruce
Subject: Re: converging work streams for all SeaNET features