Web Analytics is current context is capturing, measuring, summarizing and visualizing visitor information for understanding contribution of various pages and contents of a website in achieving its business objectives.
A list of web analytics tools such as Google Analytics, Webtrends , IBM Coremetrics, Adobe SiteCatalyst and a few others is available to help organizations in capturing the visitor information from their website. They also provide standard options of defining certain parameters, measuring visitors and summarizing & visualizing key metrics.
Web Analytics tools leverage two methods to collect data for web visitors. These are web server log files and page tagging. Both of these methods have advantages and disadvantages associated with them.
Google Analytics provides some of the below metrics for a website :-
Web analytics is different from social media analytics. In web analytics typically, we monitor and analyze visitors to a website not a social media website. Also, we have a lot more information on customer comments and interactions for social media websites.
Web Analytics inputs can be employed in a variety of ways – from improving website performance enhancement to dynamic customer engagement.
Currently, majority web analytics work revolves around deployment of the web analytics tool, and monitoring website performance. The information is certainly at a visit level but identification of a person is a bit fuzzy. The visitors can’t be linked to the existing customers unless visitors are identified (registered & logged in).
Organizations are looking to leverage web data for building predictive models and making better decisions. Additionally, the web data can augment the current predictive modeling data sources and improve accuracy rate.
Web Data for Predictive Modeling Case Study :-
On Kaggle website there was a competition forStumbleUpon. The objective of the competition was to predict or classify a web page as either "ephemeral" or "evergreen". The predictive model will be used to recommend pages to the visitors. The data available involved page attributes and also contents of the web pages. This is a case of using web data both structured and unstructured for classification problem. The teams have used range of statistical and machine learning techniques.