A question to Lebanon watchers

This may be rather naive, and considering the vitriol being thrown around on the issue of Lebanon these days I want to tread carefully: but how come analysts have such detailed knowledge of the voting patterns according to sectarian affiliation in the recent Metn by-election? Are these published in official records? Are they based on exit polls? To see what I mean, see for instance this analysis which takes to task a recent Hassan Fattah article in the NYT (as many March 14 supporters have been doing, which for Hassan must be a change from being attacked by March 8 supporters):

In 2007, Michel Aoun’s candidate won the seat by a razor-thin margin – 418 votes, or 50.2%. Overall, Metn voters were split down the middle in the by-elections. This is a far cry from Aoun’s dominant showing in 2005, where he claimed to have won 70% of the Christian vote in Mount Lebanon. In addition, Gemayel was the clear choice of the Maronite community, winning 58.6% of Maronite votes to the FPM’s 40.7%. Contrary to Mr. Fattah’s assertion that the Metn by-elections showed that the Christians are “increasingly alienated” from the March 14 coalition, the actual breakdown of results shows the exact opposite.

So, again, the question is: how is it that such detailed info on what I had assumed was a secret ballot is available, and how reliable is it?

Update: I left a comment on the NOW Lebanon site linked above asking the same question. The site's staff pointed me to the chart below, which does offer a breakdown by community. Hover on the bars to get the community information. There is no sourcing or methodology info here, however, so I'd still appreciate it if someone can confirm that these statistics come from officials and that this is routine dissemination in Lebanon.