The hosting company you select is actually the group that can collect statistics on site visits, but some of the information that should be available includes:

• # of hits (Includes html pages plus graphics on the page that are requested by the visitor. This is a grossly exaggerated number, since it includes the number of graphics on each page, so is of little use other that to inflate your perception of how successful your site is.)
• # of pages (How many pages were actually viewed. If your home page was viewed 450 times, and your second page was viewed 230 times, number of pages viewed would total 680 plus the number of times all your other pages were viewed. If one of your goals is to get visitors to dig deeper into the site, and spend some time on it, you would want to see a relatively high average # of pages viewed. To determine that, divide # of pages by # of distinct hosts, and you'll get a good approximation. Obviously you wouldn't want your average to be too close to 1 or 2.)
• # of distinct hosts (This isn't an exact count due to the use of proxy servers, but it will be an approximate count of the number of visitors came to your site.)
• # of times each page was viewed (You can find out which pages are drawing the most attention, and then experiment with placement, colors and other effects to get your audience where you want them to go.) Referrer report (Tells you where your visitors are coming from, giving the last page viewed before coming to your site. This is a way for you to test out the effectiveness of any Internet advertising dollars you may have spent, as well as evaluate which search engines and other links you've established are contributing the most to your site's traffic.
• Top-level domain of visitors (Tells you how many pages were viewed from educational institutions, government agencies, commercial entities, other countries, etc.)