Statisticus, Why and How?
Part I - Introduction
by Hakan Eskici
Contents
Abstract
1. Introduction
2. Measurement Tools
a. Page Counters
b. Web Server Logs
c. Our Own Counter
3. Simple Statistics
4. Adding More Fields
5. Data Normalization
6. Data Gets Valuable
Abstract
This article focuses on basic statistical analysis methods and the techniques to implement
them into the web pages.
1. Introduction
* A latin doctrine, which states : "rules change when circumstances change".
The idea of interactive web has been kind of derived from the supply and demand balance.
You are successful in wall street bussiness as long as you have the ability to drive
your customers expectations. By creating a false demans bias, you are stronger in your
way.
This is not the way we devevelopers work.
We, the devs and admins, should understand our visitor expectations, and give them
what they really need.
In web, we have the tools for persistent data, we have the tools to do math.
That means we have the tools to measure the behaviour of our visitors. By analyzing
the past data, we may understand the trend and guess the future demand. The
methods for forecasting is strongly quantitive and sometimes requires a good
command on mathematics. We will not be dealing with regression analysis or
curve fitting, but rather be focusing on simple methods such as moving averages
and line fitting to analyze our data.
Therefore, understanding the power of statistical analysis is important for us.
Second part of the article focuses on different
types of the measurement tools: page counters, web server logs and our own
counter.
|