Information Systems-clothes Recommendation Systems

ISOM 221 Information Systems Analysis and Design Fall 2008/2009 Team Project Final Report

Group Name: L1-07 Group Members: AU, Kai Hin Alfred CHAN, Pui Shu Grace LAU, Chor Ken Elvis LAU, Wing Ki Winnie PUN, Chun Yiu Andy TSANG, Hang Fung Napo

1

Table of Contents Project Drivers 1. The Purpose of the Product 2. Clients, Customers and other Stakeholders 3. Users of the Product

3 4 5

Project Constraints 4. Mandated Constraints 5. Naming Conventions and Definitions 6. Relevant Facts and Assumptions

9 10 10

Functional Requirements 7. The Scope of the Work 8. The Scope of the Product 9. Functional Requirements

11 12 14

Non-Functional Requirements 10. Look and Feel Requirements 11. Usability Requirements 12. Performance Requirements 13. Operational Requirements 14. Maintainability and Portability Requirements 15. Security Requirements

23 23 23 24 24 24

2

Project Drivers 1. The Purpose of the Product

The user problem When buying clothes, people sometimes do not know which clothes to choose or which clothes can best fit their clothes that they have bought before. In order to help solve this problem and enhance the shopping experience of buying clothes of customers, Clothes Recommendation System will be implemented. Background to the project effort To enable this system to work, the company has to first set up a membership club for regular customers. Individual accounts can then be established for each member to record their purchasing pattern. In the system, customers can input their personal information, like height and weight. This data can help analyze the clothes that customers need. Also, from the purchasing history, the system can learn about the preferences of the customers. Thereafter, it can make customized suggestions to members on the clothes that they may like. Goals of project Implementing this Clothes Recommendation System, company can understand customers’ preferences easier. Therefore, the company can choose to produce those favorite clothes more so as to reduce the cost of scarping the unsold inventories or the cost of storing these products. They can also improve the customer satisfaction and increase the sale volume by the customized suggested clothes, which are suitable for the customers and they are likely to buy the products. Moreover, the company can reduce 15% of costs on marketing research stages. The key functionalities of our system are handling recommendation requests and sending personal newsletters. Customers can request the clothes recommendations through the Internet or the terminals in the shops. The system will then generate the suggestions based on the information mentioned above. The personal newsletters would be sent periodically, which contain the suggested new arrivals. 3

2. Client, Customers and other Stakeholders Customer Customers are the major stakeholders of the system. After registering as a member of the clothes company, the customer can receive customized clothes recommendation through the system. They can also receive personal newsletter which is generated by the system. And the personal newsletter will contain the new arrival that match with them. Clothes Company The clothes companies would be another key stakeholder. It is because the companies can increase both the sales volume and customers’ satisfaction by implementing this system. They may also reduce 15% of costs on marketing research. Therefore, the system can create more profit for the companies. Moreover, the system also helps the company to enhance the brand name as well as to achieve a better competitive position by the customized service Sales Staff Those sales staffs are also benefited from this system. The system can improve their productivity, since they can save time to ask for necessary customers’ preference. Also the system can help the staff to improve the customers’ satisfaction because of the right clothes they can suggest to the customers. Fashion Designer The fashion designer classifies clothes in the company clothes database into different categories. The fashion designer determines the types of clothes recommended to customers by designing the matching principle for the matching system to match the clothes and the customers. In addition, having the system, they can spend less time on determining customers’ preference, and also they can produce more popular clothes.

4

3. Users of the Product List of Actors: 1. Customer 2. Matching System 3. Company Sales Database 4. Customer Database 5. Company Clothes Database 6. Purchasing History Database 7. Technicians 8. Fashion Specialist Actor Specification Actor Name: Customer Type: Primary Role Description:

Personality: I, R

Abstract: No

A customer is a person who has opened a membership account or would like to open an account. The customer is the main reason for the existence of clothes recommendation system. The customer interacts with the system to obtain recommendation through company webpage or terminals inside the shops. The customer also receive personal newsletter generated by the recommendation system. Actor Goals:

 Open membership account  Manage account information, including their preferences and basic information  Request clothes recommendation  Receive Personal Newsletter Use Cases Involved with:

 UC-001 Set up membership account 

UC-002 Retrieve customer data

 UC-003 Request recommendation 

UC-006 Send Personal Newsletter

5

Actor Specification Actor Name: Matching System Type: Secondary Personality: E, R Role Description:

Abstract: No

A matching system is a system which matches the data of the clothes with different customers. The system retrieves data from different systems and actors and sends the result to the clothes recommendation system. This system is used to match the clothes with the customers, based on their style, physical appearance, etc. Actor Goals: 

Provide the matching result to the recommendation system

Use Cases Involved with:

 UC-003 Request recommendation  UC-006 Send Personal Newsletter

Actor Specification Actor Name: Company Sales Database Type: Secondary Personality: E Role Description:

Abstract: No

A company sales database stores data about the company sales. The company sales database provided information about the popularity of the recommended clothes. This information is required because the clothes recommendation is listed in the order of the clothes popularity. Actor Goals:

 Identify the popularity of the recommended clothes Use Cases Involved with: 

UC-003 Request recommendation

6

Actor Specification Actor Name: Customer Database Type: Secondary Personality: E,R Role Description:

Abstract: No

A customer database is a database which stored the data of the customer. The customer database exists because storing the customers’ data can help the system to provide the most suitable service to them. Customer can also edit the data through the web or terminals. And the system also gets the customer’s email address from this database in order to provide personal newsletter to the customer. Actor Goals:

 Update the database  Record customers’ account  Provide data to the system, eg email address and customers’ physical characteristics Use Cases Involved with:

   

UC-001 Set up membership account UC-002 Retrieve customer data UC-003 Request recommendation UC-006 Send Personal Newsletter

Actor Specification Actor Name: Company Clothes Database Type: Secondary Personality: E, R Role Description:

Abstract: No

The company clothes database is a database that stores the clothes items that the company sells, including shirts, skirts, socks, etc. When there are new clothes, the database will be updated. This database works with Purchasing History Database to provide information for Matching System. Actor Goals:  Keep the up-to-date information on all the clothes that company has  Provide source to the matching system Use Cases Involved with: 

UC-003 Request Recommendation

 UC-004 Update Company Clothes Database  UC-006 Send Personal Newsletter

7

Actor Specification Actor Name: Purchasing History Database Type: Secondary Personality: E, R Role Description:

Abstract: No

A Customer Purchasing History Database is a database which records each and every purchasing history of an individual customer. This system is crucial for the Matching System, which compares latest stock to customer’s purchasing history, preference, advice from Faison Specialists and more. The database provides valuable sources to creating meaningful matching results for Customers. Actor Goals:

 Record purchase from Customers  Provide source to the Matching System Use Cases Involved with:

 UC-003 Request Recommendation  UC-006 Send Personal Newsletter

Actor Specification Actor Name: Technicians Type: Secondary Role Description:

Personality: E

Abstract: No

A technician is a person who is responsible for the supporting work of IT and machine related things. They include the terminals and the company webpage. He maintains the system so that the system can function normally. He also needs to update the system, both hardware and software from time to time. Actor Goals:

 Ensure the normal performance of the system Use Cases Involved with:

 UC-005 Maintain Terminals and Webpage

8

Actor Specification Actor Name: Fashion Specialist Type: Secondary Personality: E Role Description:

Abstract: No

A fashion designer is a person who updates company clothes database and determines the matching principle. The fashion designer classifies clothes in the company clothes database into different categories. Each category of clothes fits people of a particular physical characteristic. The classification of clothes allows the matching system to match clothes to customers according to their physical characteristics. Actor Goals:

 Classify clothes into different categories Use Cases Involved with:

 UC-004 Update Company Clothes Database

Project Constraints 4. Mandated Constraints 







Security is one of the constraints of the system. The clothes recommendation system has to make use of a lot of customers’ personal details in order to make clothes recommendations. Therefore, a high security has to be maintained in the system in order to protect the customers’ privacy. The security will be using SSL. Since the system requires the Company Sales Database which is created by other system within our company. The Company Sales Database must have enough sales information for our system to process the popularity of the recommended clothes. Customers should have network connections and web browsers with Java installed in their personal computers. All staffs are well-trained to teach new customers about this system.

9

5. Naming Conventions and Definitions  Company Sales Database A company sales database stores data about the company sales. This database is created by other systems inside the company. The company sales database provided information about the popularity of the recommended clothes.  Company Clothes Database The company clothes database is a database that stores the clothes items that the company sells, including shirts, skirts, socks, etc. When there are new clothes, the database will be updated. This database works with Purchasing History Database to provide information for Matching System.  Customer Database A customer database is a database which stored the data of the customer. The customer database exists because storing the customers’ data can help the system to provide the most suitable service to them. This database also provides email address for the system to generate personal newsletters.  Purchasing History Database A Customer Purchasing History Database is a database which records each and every purchasing history of an individual customer. This database is also created by other systems within our company. This database will be updated when the customers bought new clothes by other systems. This database should have enough information for our system to process the matching.

6. Relevant Facts and Assumptions 



The system can be accessed through both the Internet and the terminals in the shop. Since one way of accessing the system is web-based method, customers can login to our system anytime anywhere. If customers purchase the clothes in our system through the Internet, they can choose to pick up the clothes at our physical shops or can have a deliver service, which belongs to another customer service system inside the company.

10



The online payment in our system is provided by other trustful 3rd parties, such as HSBC. And our system is not obligated to any transaction problems.

11

Functional Requirements Functional 7. Requirements The Scope of the Work 7. The Scope of the Work

12

8. The Scope of the Product List of Use Cases: 1.

2.

3.

4.

5.

6.

7.

8.

9.

UC-001 Set up membership account New customers set up their own accounts through the Internet or terminals inside the shops. UC-002 Retrieve Customer Data Existing customers can update or change their personal data and their own preferences from time to time. UC-003 Request Recommendation Customers request personal clothes recommendations from our system. UC-003 A1 Request Recommendation Alternative for Sold Out Company does not have enough inventories for the recommended clothes that customers would like to buy. UC-003-E1 Print Receipt Discount Coupon Customers would like to print the transaction receipts or they would like to print the discount coupon for them to use next time at our physical shops. UC-004 Update Company Clothes Database Fashion specialists classify the new clothes purchased into different categories of clothes and then update the clothes into the Company Clothes Database according to the categories. UC-005 Maintain Terminals and Webpage Technicians repair the terminals and web servers when they are broken down or errors occurred. And they also make necessary updates to the hardware or software from time to time. UC-006 Send Personal Newsletter The system generates personal newsletters, which contain the recommended new arrivals and other pre-written fashion articles, twice a month. Then, system sends them to the customers’ email addresses. IUC-001 Log in and Verify Customer System handles customers’ login and verifies them by checking the password entered.

13

14

8. The Scope of the Product 9. Functional Requirements Use Case ID Use Case Actors Description Pre-conditions Flow of Events

UC-001 Set up membership account Customer[P], Customer Database[S] New customers browse webpage, enter personal data, enter login name and password and receive confirmation of registration. Webpage is shown with welcome screen 1 The system requires the customer to enter his/her personal data.

2 3

Customer enter their personal data The system checks whether the customer have registered before from customer database.

4

If the customer have registered before 4.1

The system shows the message that the customer have registered before.

4.2 Go to step 1 5

Customer database send back the message that the customer have not registered before.

6

The system requires customer to enter the log in name and password.

7 8

Customers enter the log in name and password

9

If the log in name is used

The system checks if the log in name is used Go to step 7

10 The customer database send back the message that the log in name is not used

11 The system sends the customer data to the customer database 12 The system shows the confirmation message of registration to the customer. Post-conditions Alternative flows Priority Non-Functional Requirements Assumptions Source

Use Case ID Use Case

13 The system shows the homepage The registration is accepted by the system Customer already registered, log in name used High The personal data should be encrypted. The system should be ready for people to register most of the time The customer understands the languages we use in the webpage. Questionnaires, Document Analysis

UC-002 Retrieve Customer Data

15

7. The Scope of the Product

9.1

Actors Description Pre-conditions Flow of Events

Customer[P], Customer database[S] Existing customers may change personal data, fill in types of clothes they want and preferences on clothes. Screen of editing personal information is on. 1. Include IUC-001 Log in and Verify Customer.

2. System asks if the customers want to change personal data 3. If customer wants to change previous saved data 3.1System shows previous customer’s personal data 3.2 System requires the customer to fill in what types of clothes they are seeking, like jacket, vest, sweatshirt, cardigan, tee shirt, shirt, polo shirt, jeans, pants, dress, etc

3.3Customer enters the types of clothes he wants 3.4 System requires the customer to fill in preferences the customer wants, like long-sleeved, short-sleeved, hooded, wrinkled, slim-typed, baggy, newly arrived, etc 3.5 Customer enters the preferences of clothes he wants 3.6 Customer enters other data that he would like to change.

3.7Customer confirms all the changes. 3.8System saved the new data and sends back updated data Post-conditions Alternative flows Priority Non-Functional Requirements Assumptions Source

Use Case ID Use Case Actors Description Pre-conditions Flow of Events

to Customer Database Main menu is on again / High All required information, like weight, height, clothes types and preferences, are “correct & truth” so that system can generate right recommendations later. Customer’s basic information will not have much change but their preferences on clothes will change from season to season. Market Analysis

UC-003 Request Recommendation Customer[P], Matching System[S], Customer database[S], Purchasing history Database[S], Company sales database[S], Company clothes database[S] Customer logs in, requests for clothes recommendations. System displays the suggested clothes to customer. The main menu is displayed with clothes recommendation option. 1. Include IUC-001 Log in and Verify Customer.

16

2. System displays a choice menu.

3. Customer selects for a clothes recommendation 4. Customer inputs some primary constraints, such as type of clothes.

5. System retrieves necessary data including customers’ preferred clothes types, physical characteristics, company purchased clothes from Customer Purchasing History Database and Customer Database and Company Clothes Database respectively.

6. System sends all the retrieved data to the Matching System for processing the recommendations

7. Matching System sends back the recommendation results. 8. System retrieves sales data from the Company Sales Database.

9. System calculates popularity of recommended clothes based on sales data and sorts the received recommendation results in decreasing order of popularity.

10. System displays recommended clothes. 11. If Customer inputs some additional constraints, such as the price range of the clothes.

11.1 System removes the irrelevant results from the list. 11.2 Go to step 10 12. Customer chooses clothes for purchase.

13. If the clothes are sold out, execute to UC-003-A1. 14. Customer enters the transaction details.

15. If Customer wishes to print the transaction receipt and discount coupons for next time, extend to UC-003-E1 Post-conditions Alternative Flows

16. Thank you screen is on. The main menu is displayed again. Company may not have enough inventories at that time. Customer want to print receipts or coupons

Priority Non-Functional Requirements

Customer adds more constraints while recommending clothes. High The system can display 100 recommended clothes items for one customer at one time. The system can generate the recommended clothes list within 2

Assumptions Source

seconds. The customers’ purchase records and details for their physical characteristics are available Customer Services Manual

17

Use Case ID Use Case Actors Description Insertion Point Pre-conditions Alternative Flow of Events

UC-003-A1 Request Recommendation Alternative for Sold Out Customer[P], Matching System[S], Customer database[S], Purchasing history Database[S], Company sales database[S], Company clothes database[S] The clothes that the customer has chosen are short of inventories. Customer needs to place a special order for buying the clothes. UC-003; flow 17 System notices that the chosen clothes is sold out at that time. 1. System informs customer that the clothes is out of stock 2. System asks customer of whether to reserve the clothes 3. If customer chooses to reserve the clothes 3.1 The clothes is automatically reserved 3.2 The system displays an acceptance screens and sends notification email to customer when the clothes is available 3.3 Go to step 4

Post-conditions Priority Non-Functional Requirements Assumptions Source

4. The system reminds manager to place an order for the sold-out item The main menu is displayed again. High Sold out items will be repurchased less than a week. / Customer Services Manual

18

Use Case ID Use Case Additional Actors Description Extending Use Case Extension Point Guard Condition Flow of Events

UC-003-E1 Print Receipt Discount Coupon / Allows customer to print the transaction receipt and the discount coupons for them to use next time. UC-003 Request recommendation UC-003; flow step 20 Transaction Receipt and Discount Coupon printing options are chosen and enabled. 1. The system ask the customer if he/she would like a printed transaction receipt and discount coupon

2. If customer wants a transaction receipt 2.1 Print statement 2.2 Go to step 3 3. If customer wants a discount coupon 3.1 Display different types of discounts coupons 3.2 Customer choose one desired discount coupon Post-conditions Alternative Flows Priority Non-Functional Requirements Assumptions Source

3.3 Print the chosen discount coupon Return to UC-003 and continue on flow step 21 / Medium Statement and discount coupons can be printed out within 20 seconds / Interview of customer services

19

Use Case ID Use Case Actors Description Pre-conditions Flow of Events

UC-004 Update company clothes database Company Clothes Database[S], Fashion Specialists[S] Fashion specialists classify new clothes and update the detailed information to Company Clothes Database according to the categories. Company has purchased new clothes 1. New clothes arrive 2. System sends acknowledgments to the fashion specialists.

3. Fashion specialists classify the new arrivals into different categories of clothes.

4. If the new arrivals are classified to be a new category 4.1 Fashion specialists create a new category 4.2 Acknowledgments are sent to the front-line shops 4.3 Go to step 5

5. Fashion specialists definite the suitable targets for each clothes and write a short description about the clothes.

6. Fashion specialists update those clothes with the detailed information into Company Clothes Database according to the Post-conditions Alternative Flows Priority Non-Functional Requirements Assumptions Source

categories. Those new clothes are sent to the physical stores. The new arrivals are classified to be a new category Medium System allows many fashion specialists to input the new arrivals’ information at the same time. Fashion specialist can correctly categorize all new clothes Business Operation Manual

20

Use Case ID Use Case Actors Description Pre-conditions Flow of Events

UC-005 Maintain Terminals and Webpage Technician[S] Technician repairs the terminals and web servers, and makes necessary updates to the hardware or software. Terminals or web servers occur errors or broken down 1. System sends acknowledgements to technician

2. Technician checks the problems of terminals or web servers 3. Technician repairs the terminals or web servers by replacing some hardware or reinstalling the software. 4. If the errors cannot be fixed immediately

4.1 Technician informs the head office for replacement 4.2 Technician places a notice to both staff and customers 4.3 Go to step 6

5. Technician reconfigures the terminals or web servers according to the latest company’s policies

6. Technician documents the errors details and sends a copy of the documentation to the head office for reference.

7. Technician allow customer to use the terminals or to access the Post-conditions Alternative Flows Priority Non-Functional Requirements Assumptions Source

web servers again. Terminals or web servers are repaired and updated Errors cannot be fixed immediately. Medium The system can send acknowledgements to technicians within an hour when errors occurred Technician performs system maintenance monthly Interview, Document Analysis

21

Use Case ID Use Case Actors Description Pre-conditions Flow of Events

UC-006 Send Personal Newsletter Customer[P], Customer Database[S], Matching System[S], Company Clothes Database[S], Purchasing History Database[S] Bi-weekly Newsletter is distributed to customer. Content of newsletter includes personalized recommendation and pre-written fashion articles. System meets the scheduled time to distribute the newsletter 1. System retrieves mailing list of customers who have subscribed to newsletter.

2. System retrieves customer data and new arrivals information for matching

from

Customer

Purchasing

History

Database,

Customer Database and Company Clothes Database. 3. System sends all the retrieved data to Matching System 4. Matching System sends back recommendation results. 5. Recommendation

results

are

embedded

into

pre-written

newsletter templates. Post-conditions Alternative flows Priority Non-Functional Requirements

6. System prepares and distribute newsletter to customers. System has sent the emails to customers. / Medium The system prepares newsletter at least bi-weekly. The appearance and the content of the newsletter should be attractive to customers as a way to promote the store instead of

Assumptions

spamming customers. Customers have email address for receiving newsletter. Customers have set up accounts and input necessary information

Source

for the recommendation process to be feasible. Customer Services Manual, interviews with managers

22

Use Case ID Use Case Description Including Use Cases Pre-conditions Alternative Flow of Events

IUC-001 Log in and Verify Customer Customer logs in; system verifies the customer and checks the password. UC-002 Retrieve Customer Data; UC-003 Request recommendation System notices that customer entered into webpage or terminals. 1. System asks customer to input account number and password. 2. Customer enters account number and password. 3. If account number does not exist or password is incorrect 3.1 If customer enters incorrect password 5 times 3.1.1.Customer account is locked for an hour. 3.1.2.The process ends and main menu displays again. 3.2 System displays error message. 3.3 Go back to step 1

Post-conditions Alternative Flows Priority Non-Functional Requirements Assumptions Source

4. System displays the welcome back message. System is back on the next flow step right after this included use case was invoked. Customer inputs wrong account numbers or passwords High The system can verify the customer for login within 2 seconds. Customer can understand the language we used. Operational Procedures Manual

23

NON-FUNCTIONAL REQUIREMENTS 10. Look and Feel Requirements 10.1 Interface: The system should be simple and user-friendly. It should contain clear instructions to guide users. This shortens the time for the customers and staff to learn how to control the system. 10.2 Style: The system should have a stylish design. This helps to create a more fashionable image of the company. 11. Usability Requirements 11.1 Ease of use/learning: The system must be simple. People with little computer background should also be able to operate the system without difficulties. This allows the system to serve a greater range of customers 11.2 Personalization: The system can provide customized clothes recommendation to customers. The selection criteria will be based on the customers’ purchase history, their physical characteristics and preferences. 11.3 Access considerations: The online purchasing system can only be accessed by registered members and administrators. This ensures that the personal information of the customers is protected. 12. Performance Requirements 12.1 Speed: The system can update all databases real time so that both customers can staff can get the most updated information from the databases. The Matching system should be able to give clothes recommendation to users within 2 seconds. 12.2 Availability: The system must be available 24 hours through the internet; so that the users can order the recommended clothes at any times that they want.

24

13. Operational Requirements 13.1 Technical Environment: The system must be able to retrieve and integrate data from various databases. This is because the clothes recommendation is made based on many different data (e.g. customers’ preference, purchasing history). Therefore the system should be able to process a large amount of data. 13.2 Physical Environment: The system must be accessible through web browsers and the terminals in the stores. This allows users to assess to the system more easily and brings them convenience. 14. Maintainability and Portability Requirements Regular maintenance would be carried out by the technicians. Regular maintenance can keep the system in a good repair so that when the customer need to access through the machine or the Internet and the administrators need to retrieve data from the customer, they can do in a very smooth way. Another advantage would be that the data can be kept well and would not be lost. We have a disaster recovery plan. If the servers are damaged and data cannot recover, we will use the backup data which is stored separately to recover the lost. 15. Security Requirements 15.1 Access: Customers would have their own login name and password. They can use their login name and password to access their own account but not the others. Only the administrators have the right to access the customers’ data and are authorized to login into the database. 15.2 Privacy: SSL would be used when data are being transferred over the Internet in order to prevent unauthorized people to retrieve the data so that customers’ privacy can be protected.

25