How to the predict the number of daily visitors in six months' time based on the past three? I do not know anything about statistics. I am a software engineer.
I had a question in my mind and I think here is the place to get my answer.
Suppose that I have data for 3 months. The data shows my visitors on a daily basis. Is there a possible way to know how many visitors I will have in e.g. 6 months?
If yes, I guess there are some factors to be considered. What are they?
Also, the number of visitors in the 180 day will be predicted only approximately and must have some risk on it. How can the risk be defined?
Any links or info about this are very welcome.
Thank you community.
 A: Note: the original question title was "Is there a way to predict future change based on data?" This answer is phrased as a response to that
The short answer to your question is "yes, with an appropriate model".
The long answer is "no, probably not", because what constitutes "an appropriate model" is highly dependent on your data, and on any assumptions you make about that data.
For example, do you envisage linear growth (i.e. the number of visitors increases by approximately the same number each month), or exponential growth (visitor numbers increase by a similar percentage each month)? Both of these might be optimistic, it might be that you are experiencing annual periodic visitations, and are just seeing the upswing in your three months of data, in which case you might expect a downswing of equivalent magnitude three months later. You might equally see spikes in your data (very common on websites), due to external factors that are outside your control, and largely unpredictable (e.g. you got re-tweeted). 
Lots of the above can be modelled by various statistical methods, such as an ARIMA model, which will separated out the trend and cyclical components, so that you can see more clearly what's going on. Whether models like this are appropriate for forecasting is highly dependent on the quality, length, and various other attributes (like variability) of your data. There's a good introduction to this kind of stuff, with examples in R, at http://a-little-book-of-r-for-time-series.readthedocs.org/en/latest/src/timeseries.html
Probably a better way to model your site's visitations in many cases is to base your model on some real world factors. For example, if your website is for a summer clothing company, or a sports manufacturer (say, soccer balls), you really ARE going to see annual periodic patterns in your data, as those products are only in high demand during half the year. If your website is for a restaurant, you'll probably find that you have a weekly periodic pattern. An ARIMA model will pick up these cycles if your data set is long enough. 
If your website is promoting a Thneed (or some other new product that everyone needs), you might be able to make the assumption that website growth will be exponential, just because the product is so brilliant, or whatever. You could then fit an exponential trend to the data, and see where it takes you. This assumption would have to be defended (and in most cases that would be pretty difficult). This emphasises the point that, especially when used for prediction, statistics isn't just about putting numbers on things, it's about making good arguments, any using numbers to back yourself up.
For your specific case, it's a bit hard to give any useful advice without seeing the dataset, but if the set is sparse, the answer would have to be "don't bother" (in particular, I think three months of data is highly unlikely to be adequate for predicting six months ahead - usually you're gonna want a much larger sample than your prediction interval).
Ultimately, you're probably better off putting effort into site design, promotion, and more importantly, quality content, because this is what's going to decide how the future pans out for your website. You never know, if you spend a week learning the stats required for this prediction, you might miss adding the one piece of timely content that gets slash-dotted, and shoots you into the big league. On the other hand, stats is worth learning for other reasons, like being able to cut through other people's bullshit, and a week out of your life learning some basic stats is probably time well spent.
