共计 3398 个字符，预计需要花费 9 分钟才能阅读完成。
CS209A Project, Spring 2019
Description
With the increasing popularity of online shopping, it has brought with massive online
consumers and the rapid growth of data on merchandise information. Many large-scale ecommerce
system or online-shopping sites provide functions for commodity evaluation. A
customer can make a review on a product or service that he has purchased and used, or had
experience with.
Before buying a product or service, many customers often look at the quantity of a product
sold and its evaluating stars, and may also carefully read the evaluating comments made by
other customers. So, customer reviews are very important for evaluating commodities.
The goal of the project is data analysis based on customer reviews of some commodities
sold online. You are asked to complish the following tasks:
(1) Collect customer reviews of the corresponding commodities on seleted commercial
websites;
(2) Preprocess the collected data, including deleting the duplicate data and short trivil
reviews, segmenting words and calculating word frequencies.
(3) Show the word segmentation and word frequency statistics on GUI created with JavaFX.
Provide an easy way to select some word as classification markers of a certain class.
(4) Classify and quantify the grade of target commodities.
(5) Store the evaluation results of target commodities and update the commodities
evaluation periodically according to the setting period.
Project Requirements
(1) You are asked to write a JavaFX graphical user interface displaying your information,
including the word frequency statistics and the grades of target commodities and so on.
You should also design an intuitive way to guide a user to choose some words as
classification markers.
(2) You are asked to write a Web Crawling Program to collect customer reviews and update
periodically.
(3) The number of commodities should be at least 3. You may analyze more commodities by
category, but the number of categories should be at least 3 and you should also analyze
the correlation among them. Correspondingly, you could get higher scores for category
analysis instead of individual commodity analysis.
(4) You should store the original data, the evaluation results and responds classification
markers into a .json or .xml file(or files). When open your application next time, your last
analysis results will be displayed on screen.
(5) Every team should prepare a PPT to explain how you process the data and how to do the
data visualization. In your PPT, you should introduce your team members with their
contribution rate for each member.
CS209A Project, Spring 2019
2 / 2
Task 1. Data Collecting (20%)
You are asked to write a Web Crawling Program to collect customer reviews and update
periodically from e-commerce websites. JD, Taobao, Amazon, dianping, meituan are good
choices. But PAY ATTENTION that that most of EC websites have ANTI-CRAWLER technique,
so don’t crawl too much data at one time.
Task 2. Data processing (35%)
You are asked to process the data that you crawled.
(1) Preprocessing, including deleting duplicate data and short trivil reviews.
(2) Word segmentation. ICTCLAS is recommended for word segmentation.
(3) You should select some reasonable classification markers and tell why in your PPT.
(4) Design a reasonable method to grade the commodities. If you analyze commodities by
category, you should not only analyze the commodity itself, but also the relationship
between categories.
Task 3. Data visualization (35%)
(1) At least a TableView, filtering on three different criteria.
(2) At least two charts or an advanced chart.
(3) Animation is a better option.
Here is an excellent example:
https://www.washingtonpost.co…
m=.5a6417cd90f2
Some other examples, please see the images and the animations in the same zip with
specification.
Task 4. Data Persistence (10%)
You are asked to store the original data, the evaluation results and responds classification
markers into a .json or .xml file(or files).
Reference
Be creative and have fun!
