It depends on what you want your users to learn: For general overviews, a cloropleth map. If you want to show quantitative values with high accuracy for comparison, a sorted bar chart allows the eye to judge line lengths to see volume.
Do you want users to see what the general popular locations users are visiting from? Or is volume accuracy per location not relevant? (I would think it is, but please clarify)
You can use a Cloropleth map, in which volume is represented by color intensity:
viewing web traffic by specific points; bubbles beware.
If you're relying on a map with volume data, a lot of visualizations involve maps with bubbles for specific locations; cities for example. (proceed with caution if comparing volume with accuracy is important):
Bubble charts on a map can show general volume and location at a glance, but circle area sizes are hard to parse for accuracy.
Try a sorted bar if the map representation is not as important (I don't know the full details of your use case). Here's an example of horizontal bars sorted by numbers:
Users can scan from top to bottom, and compare values. You can also allow users to drill into a country and sort users by states or cities.
If you have room, you can have both a map and a sorted bar (or a sorted data table).
If real estate is not a concern, you can show a high level map, followed by a chart or table in sorted order. As you drill into the map, you can update the data below it:
This allows you to show a country w/ state (and then city) drilldown.
TL;DR Please do not use a pie chart
Here's an article(pdf) from Stephen Few (prolific author on data visualizations) that dives more deeply into some of the problems of pie charts: Parsing volume with slices and angular shapes:
We make angle judgments when we read a pie chart, but we don’t judge angles very
well. These judgments are biased; we underestimate acute angles (angles less than
90°) and overestimate obtuse angles (angles greater than 90°). Also, angles with
horizontal bisectors (when the line dividing the angle in two is horizontal) appear
larger than angles with vertical bisectors.
If you have countries (over 193 of them!) to compare, the amount of slices quickly becomes quite hard to parse.