<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<tt>A further update on this work:</tt><tt><br>
</tt>
<ul>
<li><tt>I found more yet bizarre phone-related tags "phone:1",
"telephone" and the like. These have all been tidied. My osmfilter
now looks like this: --keep="contact:*=* or phone*=* or
Phone*=* or alt_phone=* or fax*=* or tty*=*" Additional
suggestions for something to search on are welcome so I get
all phone numbers.</tt></li>
<li><tt>I found there were some formats used very regionally eg.
Edmonton Schools used one format consistently and Ottawa
Schools used a different format consistently.<br>
</tt></li>
<li><tt>The canada.poly filter I have been using includes Saint
Pierre and Miquelon (which does not use North American dialing
plan), as well as a few US entries (especially relations which
go near the border). If anyone knows of a canada.poly that is
tighter, can you point me in the direction? I am generally
leaving non-Canadian entries alone, but they do count in the
stats below.<br>
</tt></li>
<li><tt>There are now 67 unique tag/phone number format
combinations (down from 400+ originally) when using egrep -i
'k="[a-z:]*(phone|fax|tty)[a-z:]*" ' $OSMFILENAME | cut -d\"
-f2,4 | sed -e 's/[0-9]/#/g' | sed -e 's/[A-Z]/A/g' | sed -e
's/([a-zA-Z -]*)/(...)/g' | sort | uniq -c | sort -nr | wc -l</tt></li>
<li><tt>The bulk of the work remaining now is to reformat the big
groups of numbers that do not begin with "+1"</tt><tt>. I
will make changes by area code to limit the number of
canada-wide changesets.<br>
</tt></li>
</ul>
<tt><br>
As always, comments welcome.<br>
<br>
</tt><tt>Here is the new "top 20"</tt><tt> as of ~10am ET today:<br>
</tt><tt><br>
</tt><tt> 12555 phone"+#-###-###-####</tt><tt><br>
</tt><tt> 4453 phone"+# ###-###-####</tt><tt><br>
</tt><tt> 4060 phone"###-###-####</tt><tt><br>
</tt><tt> 3749 phone"+# ### ### ####</tt><tt><br>
</tt><tt> 2624 phone"+# ### ###-####</tt><tt><br>
</tt><tt> 2239 phone"(###) ###-####</tt><tt><br>
</tt><tt> 1292 fax"+#-###-###-####</tt><tt><br>
</tt><tt> 1032 phone"##########</tt><tt><br>
</tt><tt> 941 contact:phone"+#-###-###-####</tt><tt><br>
</tt><tt> 323 phone"+###########</tt><tt><br>
</tt><tt> 322 phone"+# ### #######</tt><tt><br>
</tt><tt> 158 contact:fax"+#-###-###-####</tt><tt><br>
</tt><tt> 117 phone:tollfree"+#-###-###-####</tt><tt><br>
</tt><tt> 109 phone"###-####</tt><tt><br>
</tt><tt> 39 phone"+#-###-###-####;+#-###-###-####</tt><tt><br>
</tt><tt> 25 phone"+#-###-###-AAAA</tt><tt><br>
</tt><tt> 23 phone"+#-###-###-####x###</tt><tt><br>
</tt><tt> 17 phone"+# (###) ###-####</tt><tt><br>
</tt><tt> 14 phone"+#-###-###-####x####</tt><tt><br>
</tt><tt> 9 phone"+#-###-###-####x#</tt><tt><br>
</tt><br>
<br>
<br>
<div class="moz-cite-prefix">On 2018-02-04 11:49 PM, OSM Volunteer
stevea wrote:<br>
</div>
<blockquote type="cite"
cite="mid:B9DF042A-3611-4795-AD46-C4185C3940F5@softworkers.com">
<pre wrap="">On Feb 4, 2018, at 8:37 PM, Matthew Darwin <a class="moz-txt-link-rfc2396E" href="mailto:matthew@mdarwin.ca"><matthew@mdarwin.ca></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Just an update on this activity.
</pre>
</blockquote>
<pre wrap="">
Again, nice work!
</pre>
<blockquote type="cite">
<pre wrap="">Here are the top 20 tags as of ~4pm ET Sunday:
10669 phone"+#-###-###-####
4392 phone"+# ###-###-####
4206 phone"###-###-####
2970 phone"+# ### ### ####
2540 phone"+# ### ###-####
2451 phone"(###) ###-####
1076 phone"##########
659 phone"+# ### #######
547 fax"+#-###-###-####
522 contact:phone"+#-###-###-####
516 phone"+###########
456 phone"#-###-###-####
446 phone"### ### ####
378 fax"+# ###-###-####
283 contact:phone"### ###-####
260 phone"+# (###) ###-####
200 fax"+###########
186 phone"### ###-####
170 phone"(###)###-####
162 fax"+# ### ###-####
</pre>
</blockquote>
<pre wrap="">
I'd appreciate others to chime in about this, but it seems where dashes and space characters overlap (are the only difference in format), those can be conflated together. I'm not sure whether dash or space ends up as "the winner," but this should reduce the number of categories.
As you consider additional conflations, you may be able to do this again and again, getting it down to a fairly small number of formats. I urge additional feedback (here would be good) before additional conflations, but (I keep saying it): nice work.
SteveA</pre>
</blockquote>
<br>
</body>
</html>