Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Generate better coverage data #14

Open
migurski opened this issue May 4, 2016 · 29 comments
Open

Generate better coverage data #14

migurski opened this issue May 4, 2016 · 29 comments

Comments

@migurski
Copy link
Member

migurski commented May 4, 2016

Companies like Google and HERE produce coverage summaries of their data:

We’ve got a visual version of this in the maps at https://results.openaddresses.io, but a tabular form might be useful as well.

@iandees
Copy link
Member

iandees commented Feb 27, 2017

@migurski
Copy link
Member Author

That Mapbox map is in dire need of visual design help.

@migurski
Copy link
Member Author

migurski commented Mar 9, 2017

Making some progress on this in openaddresses/population repo, starting with per-country area and population estimates and writing to a useful database that can drive a coverage page:

EU

NA

@iandees
Copy link
Member

iandees commented Mar 9, 2017

I <3 those colors.

@nvkelso
Copy link
Member

nvkelso commented Mar 9, 2017 via email

@migurski
Copy link
Member Author

I have a temporary representation of this data here: http://ec2-54-89-13-7.compute-1.amazonaws.com

Feedback at this stage would be really helpful!

@nvkelso
Copy link
Member

nvkelso commented Mar 14, 2017 via email

@nvkelso
Copy link
Member

nvkelso commented Mar 14, 2017

The text in the intro helps explain the 3 parts of the table, but if you breeze over that it's not clear why the table parts are grouped that way. Might also add a column for the overall coverage grade (full, substantial, minimal)? This would allow future sorting of the table, too.

migurski added a commit to openaddresses/population that referenced this issue Mar 14, 2017
@migurski
Copy link
Member Author

Good ideas, I incorporated them all.

@iandees
Copy link
Member

iandees commented Mar 14, 2017

This looks great, thanks for putting it together! The columns are a tiny bit confusing, though. maybe the "population in land area" column should be "population covered" or "estimated population covered"?

migurski added a commit to openaddresses/population that referenced this issue Mar 14, 2017
@migurski
Copy link
Member Author

Okay, made that change. I experimented with splitting up the columns a bit, curious if this makes sense.

Old:

screen shot 2017-03-14 at 2 26 39 pm

New:

screen shot 2017-03-14 at 2 26 05 pm

@nvkelso
Copy link
Member

nvkelso commented Mar 14, 2017

Is there special HTML formatting that would allow these to copy into OpenOffice as a table? Right now I'd need to do processing to get it into a full table (as Land Area Covered, Population Covered, and Address Density all have sub cells that copy into 1 generic OpenOffice cell):

screen shot 2017-03-14 at 14 49 54

So instead A-G it'd be 3 more for A-J.

@migurski
Copy link
Member Author

I was thinking of this with the "new" variant from the comment above, with separate columns for the parenthesized parts. I'll push it live; tell me if it works for you.

@migurski
Copy link
Member Author

Results when pasting into Numbers:

screen shot 2017-03-14 at 3 07 56 pm

@nvkelso
Copy link
Member

nvkelso commented Mar 14, 2017

YES!

screen shot 2017-03-14 at 15 16 34

@migurski
Copy link
Member Author

✨ Spreadsheets! ✨

@nvkelso
Copy link
Member

nvkelso commented Mar 14, 2017 via email

@nvkelso
Copy link
Member

nvkelso commented Mar 14, 2017

The paragraphs at the top serve as legend, but are missing two items:

  • Coverage Group – move this out of the intro text.
  • Population Coverage – which summarize and link people to an new anchor tag below the table for full info.

legend-markup

@migurski
Copy link
Member Author

After some good brainstorming with @sbma44 and Michael Steffen, I’m going to dumb down the address density number a little. It’s currently the average of densities (with standard deviation) and I’m going to simplify it to the average density so the arithmetic is easier to understand.

@michaelsteffen
Copy link

@migurski you can call me by my github name, it's OK :).

👋 @iandees @nvkelso

@NelsonMinar
Copy link
Contributor

NelsonMinar commented Mar 15, 2017

Loving this work. My requests:

  1. Can't wait for a map view like the preview you teased up above! Maybe color by coverage group?
  2. To facilitate scanning, please right align all numbers. (Or align on the decimal point, but a trailing zero is probably easier.)
  3. Also for scanning, don't mix thousands and millions. Ie instead of 563K and 17M people make it 563k and 17000k. (or 0.56M and 17.00M)
  4. The Palantino font's use of descending numerals and inconsistent numeral heights makes this really hard to read. Maybe this is worse in Windows, see screenshot attached

migurski added a commit to openaddresses/population that referenced this issue Mar 15, 2017
@migurski
Copy link
Member Author

Good suggestions, @NelsonMinar. I’ve applied them all except the font thing. Can you tell what font is actually being rendered on Windows? It’s not Palatino, which looks like this for me:

screen shot 2017-03-15 at 10 26 49 am

@NelsonMinar
Copy link
Contributor

You're right, I think it's actually Georgia that my sample is showing (second option in the CSS for body). My Windows 10 system doesn't have a font named "Palatino", but it does have "Palatino Linotype" which seems to work in the CSS and looks OK.

@migurski
Copy link
Member Author

This is real now: http://results.openaddresses.io/coverage

@nvkelso
Copy link
Member

nvkelso commented Mar 22, 2017 via email

@iandees
Copy link
Member

iandees commented Mar 22, 2017

👏 This is really excellent. I tweeted about it here: https://twitter.com/openaddr/status/844666719209750528

Can we add a map like the green/red one on https://results.openaddresses.io/ but coloring countries based on population coverage percentage?

@migurski
Copy link
Member Author

Yes, absolutely! I probably shouldn't use the same colors for this.

@jharpster
Copy link

These coverage metrics are the best thing ever. Also wondering if raw counts of addresses by country are maintained some place or if that would be useful to anyone besides me.

@migurski
Copy link
Member Author

That sounds pretty useful. Maybe a thing to put on a linked page for each country, as @michaelsteffen has suggested.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

6 participants