EDA

More EDA can be found on shiny application

There are in fact two different types of data: Raw Data and Summary Data. To do EDA on the data, we do seperately on the Raw Data and Summary Data

Raw Data

Here we explore the browser distribution. Most of the visitors used chrome browser. The most used browsers are chrome, safari, firefox.

Here we explore the device category distribution. The most used device is desktop.

Here we explore the device operating system. Most of the visitors used Mac. The most used operating systems are Mac, Windows, IOS, Android.

Now we check these features correlation. The most frequent combiniation of the visits is using Chrome browser from Mac.

The most frequent combiniation of the visits is using Chrome browser on desktop.

The state with the largest number of visit is California, and the state with the second largest visit is New York.

Consistent with the number of visit, the state with the highest transaction revenue is California, and the state with the second highest transaction revenue is New York.

Now we show the distribution of visit per state, considering only visits with non-zero transactions. The total numbers in each state are smaller, and the top 2 states with largest number of visits are California and New York.

We plot the time series for transaction revenues. April has the highest revenue.

Revenue is higher in weekdays than weekends.

Summary Data

Let’s wait and have a sight of what good are people buying

Then Let’s describe some detail of the transaction behavior. Transaction won’t happen when people are viewing the contents

However, lot’s of people do exit at the content page

Also there will be no transaction at entrance page

Which suggest that people just viewing content after entrance and exit then can not make revenue

Which also suggest there should be some tag that is about transaction. Yes! eventInfo.eventAction is discribing people’s action in the website

Is social referal important for transaction? Yes, but referal seems to make people not to buy. But may just beacuse of variance