# Last edited on 2014-09-26 11:26:09 by stolfilocal

ANALYZING THE BITCOINTALK POSTS BY TOPIC

  Copy-paste the "recent posts" page to a file: https://bitcointalk.org/index.php?action=unread
  
  Manually remove the header and trailer lines (egrep won't work with UTF-8 Chinese, unfortunately)
  
  Use Emacs replace to (in order)
     
     remove lines '\012by [- ._A-Za-z0-9]* *'
     replace '|' by '-'
     replace '#' by '%'
     replace TAB by ' | '
     replace Chinese text with '[CN]<translation>'
     replace '!' by '.'
     replace ' in \(.*\) in ' by ' i@n \1 in ' (repeat)
     replace 'New «[ .0-9]*[ All]*»' by ' | '
     replace ' New in ' by ' | in '
     delete ' [|] *Last post [A-Za-z0-9: ]*'
     delete '^ |  | '
     replace '| *in \([^|]*\) *|' by '| \1 |'
     Replace funny UTF-8 characters by with 2 or 3 ascii chars
     
     change type to iso-latin-1, replace bad chars, save.
     
     Pipe through 'txtable-reformat | sort'
     replace "i@n" by "in"
     
     Remove the last two columns (numbers): replace '|[ 0-9]+|[ 0-9]+$' by ''
     Rearrange colums
     Add column with category in front
     Sort and count
   
  TO SORT UTF-8: C-x RET c utf-8 RET C-u M-| <command>

2014-06-21    ! 2014-07-25    ! 2014-09-25    !
--------------+---------------+---------------+
Posts !     % ! Posts !     % ! Posts !     % ! Category                       
------+-------+-------+-------+-------+-------+--------------------------------
    8 |   6.7 |    14 |  12.6 |    28 |  11.7 | Off-topic                      
    4 |   3.3 |     7 |   6.3 |    15 |   6.3 | Gambling                       
   18 |  15.0 |    12 |  10.8 |    58 |  24.3 | Bitcoin mining                 
   37 |  30.8 |    22 |  19.8 |    23 |   9.6 | Bitcoin non-mining             
   53 |  44.2 |    56 |  50.5 |   115 |  48.1 | Altcoins (including mining)    
------+-------+-------+-------+-------+-------+--------------------------------
  120 | 100.0 |   111 | 100.0 |   239 | 100.0 | TOTAL                          


  To (re)compute percentages, make a scratch table with lines in the format 
  "{COUNT} | 0.0 | {CATEGORY}" and pipe throught the following
  command (with the proper value of {ntot}):
  
  gawk -v ntot=120 'BEGIN { FS = "|"; nt = 0; } /^ *[0-9]/ { n = xb($1); t = xb($3); printf "%4d | %4.1f | %s\n", n, 100.0*n/ntot, t; nt += n; next; } //{ print; next; } END { printf "total %d posts (should be %d)\n", nt, ntot > "/dev/stderr"; } function xb(x){ gsub(/^[ ]+/,"",x); gsub(/[ ]+$/, "", x); return x; }'