For this lab you will implement a program that mimics data analysis for an e-commerce site. You will implement a set of classes (as determined by you) to represent the data stored in a "log" file. You will then work with the data to compute various statistics on the collected data.
Understand the provided base code that reads data from a file into an array of strings
Write code to process the data into appropriate data structures (including writing your own classes and creating hashmaps to store lists of those classes as appropriate)
Write code that operates on the data in order to compute statistics about that data (including relating the various peices of data to one another and computing averages, differences, counting given specific instructions)
Retrieve the files provided for this lab here:
These given files include base code to help in the task of analyzing log data. Your final program will need to read in e-commerce like data from a file and compute statistics based on the information in the log. The base code implements the file reading (but not the processing of the data read). You will need to complete the processing of the data for this lab, (including putting the data into appropriate data structures, i.e., not an array of strings). You are encouraged to read through the provided code.
Examine the provided log files. For example, look at the contents of
small.log. You will find various entries
representing customer use of an imaginary e-commerce site. Each entry
will appear on a single line in the file and consist of an entry
type tag followed by the corresponding entry attributes. The
different types of entries and their attributes are as follows.
START sessionId customerId
VIEW sessionId productId price
BUY sessionId productId price quantity
END sessionId
Each sessionId is a unique identifier represented as a
String,
each customerId is a unique identifier represented as a
String,
each productId is a unique identifier represented as a
String,
each price is an integer number of cents, and
each quantity is an integer.
Your program will take the name of a log file as a command-line argument. It must then read the file and store the entries for processing. Your program will then output the results of computing the statistics discussed below.
You must determine how to represent the entries and which data structures to use for processing the data. Look at the provided code, there are several clues. Though you might be tempted to simplify the data (since not all of it is used in the required analysis), avoid that temptation for now. Define classes, as appropriate, to represent the entries (give consideration for the cohesion and coupling design principles).
You should note that within the provided source file there is an example of such processing and the use of a data structure (a map) to store some of the information. Based on the material discussed thus far, this data structure is used by first creating it and then passing it to another method to be populated. This then allows for the use of the populated data structure after the file processing. Some of you may not care for this mutation-based style, which is fine, but this is the most direct approach at this time.
Your program is only expected to properly analyze logically organized log
files. In particular, for a given sessionId, all VIEW
and BUY entries will come after a START entry and
come before, if present, an END entry (it is conceivable that
a session may not have ended when the snapshot of the log is considered).
Your program must compute the following statistics.
Print the average number of items viewed by visitors (as defined
by session id) that do not make a purchase. In other words, over
all sessionIds where a purchase is not made,
compute the average number of views (VIEW entries). Note that
viewing the same item more than once should be counted - you are
computing the average 'views' not the average number of items
viewed.
For each sessionId associated with a purchase,
print, for each productId, the purchase price (as cents)
minus the average price of the items viewed during that session.
(Keep in mind that there may be multiple purchases in a given session.)
Note that you do not need to account for the quantity of the
purchased item, just compute the difference of the amount spent on a
given product.
For each customer, for each product that customer purchased, print the number of sessions in which that customer viewed that product. Note you are computing the number of sessions in which the purchased item was viewed, not the total number of views of that item.
The output of your program when run on the provided small.log
file should be as follows (the order within each breakdown is not
significant).
Average Views without Purchase: 3.0 Price Difference for Purchased Product by Session session4 product1 -100.0 session2 product1 -25.0 product3 175.0 Number of Views for Purchased Product by Customer customer2 product1 1 customer1 product1 2 product3 2
This is a longer lab and requires you to solve some data processing challenges on your own. You might want to consider these steps.
To start, make sure you understand the log file format.
Consider class designs to store important entries from the log file.
Look at main and how the data processing is proceeding (i.e. main calls populateDataStructures, which calls processFile, which calls processLine, which then processes the entries differently (you will write two of them)).
Think about the collection data structure you will use to store all the data - you will allocate that data structure in main and then pass it to the relevant methods mentioned above until you get down to the point of actually creating objects and storing them. Get a grasp of the control flow and make sure the general idea of how you are going to store data makes sense - if it does not at this point, stop and talk to your instructor or neighbor.
Fill in the important data processing steps (processViewEntry and processBuyEntry)
Stop here and see if you can now uncomment the printOutExample method and have it print out your processed data. It should look as follows for small.log:
customer2 in session4 looked at product3 in session5 looked at product1 looked at product3 looked at product1 looked at product4 customer1 in session1 looked at product1 looked at product2 looked at product3 in session2 looked at product1 looked at product1 looked at product1 looked at product4 in session3 looked at product3 looked at product2
Now you are ready to write the code to compute statistics on your data; take each task one at a time. Write out pseudo-code first in your comments. What data do you need to iterate over for each step in the computation? Then replace your logic with actual Java syntax.
Now test to make sure your output matches and you are done!
This lab is due by the end of lab in 1.5 weeks (either Tuesday 10/10 or Wednesday 10/11, depending on your section). Be sure to commit the code by the due date. Your instructor reserves the right to run further tests on your code. Demonstrate your working program to your instructor. Be prepared to show your source code. (Partial credit available up to instrutor discretion).