I use to say "Don't parse HTML with regular expressions"...
But sometime, I just do this way, because:
- Using a full blown HTML parser might be overkill;
- The HTML page doesn't change, or is generated in a consistent, predictable way.
- It is fast and convenient...

So here is my solution:
Code:

import java.util.regex.*;

String page =
"<p>There are</p>" +
"<p id='total_count' style='position: relative; top: 8px'>101,986</p>" +
"<p style='position: relative; top: 10px ;line-height: 70px  ;'>things on the site</p>";
String regex = "id='total_count'.*?>([\\d,]+)</p>";

String value = null;
Matcher m = Pattern.compile(regex).matcher(page);
if (m.find())
{
   value = m.group(1);
}
println(value); 

I replaced the " with ' to avoid backslashing them. You have to replace ' with \" in the regular expression to work in your case.
Some adjustments might be necessary, if sometime the value has no comma and/or no decimal part for example.

[EDIT] Just understood comma is thousand separator, not decimal one! So I improved the expression to handle any number of commas...

regex == 'regular expressions'.

Google that, then take a deep breath.  I'm still far from expert on them and to me pattern matches often look like gobbledigook, but there are plenty of resources online that can help you define the pattern you need to match...

Hmm, one thing, I need the applet to keep reading the html to see if the number I am grabbing has changed. The example PhiLo gave finds the figure ok, but when I tried replacing 'String page = stuff....' with a loadStrings function from another sketch, but I think this method doesnt quite work with an array as I get an error stating cannot convert string[] to String.

- how can I import the html in a format that can be read by the regex function?

Code:

  String url = "http://cagd.leedsmet.ac.uk/";
  String htmlTextlong = loadStrings(url);
  String[] htmlText = loadStrings(url);    
  String htmlTextJoined = join(htmlText, " ");
  
   int start = htmlTextJoined.indexOf("relative; top: 8px\">"); 
  int end = htmlTextJoined.indexOf("</p>", start);
  String selectedText = htmlTextJoined.substring(start, end);
  
 //println(htmlText);
String page = htmlTextlong;//instead of string page = "html stuff...."