You are on page 1of 31

ISOM 2010

Introduction to Information Systems

Lecture 13: Big Data (Analytics for Text and Network)

ALL RIGHTS RESERVED © 2020 Zhitao Yin


Lecture Roadmap



Last Class – Case Analysis

• Increasing interests in AI [machine behavior]

• AI history [Why AI Summer?]

• Machine learning [false positive, false negative, pipeline]


Today’s Game Plan

• Text Mining

• Network analysis
Decision Making of Opening a Business

Is this a good idea to open a Chinese restaurant in Las Vegas?


Market Survey Questions

Market Survey Questions

1. What is the average customer’s attitude toward Chinese restaurants in Las Vegas?

2. What are the most commonly used words in negative Chinese restaurant reviews?

3. What are the topics when customers talk about Chinese restaurants in Las Vegas?
Market Survey Questions

Collect Yelp review data to make evidence-based decisions.

Market Survey Questions

1. What is the average customer’s attitude toward Chinese restaurants in Las Vegas?

2. What are the most commonly used words in negative Chinese restaurant reviews?

3. What are the topics when customers talk about Chinese restaurants in Las Vegas?
Data Collection from Yelp for All Restaurants in Las Vegas
Market Survey Questions
1. What is the average customer’s attitude toward Chinese restaurants in Las Vegas?
2. What are the most commonly used words in negative Chinese restaurant reviews?
3. What are the topics when customers talk about Chinese restaurants in Las Vegas?
Text mining

Rating info.
Questions # 1 – Summary Statistics Calculation
Q1: What is the average customer’s attitude toward Chinese restaurants in Las Vegas?

What is your takeaway?


Questions # 2 – Sentiment Analysis via Text Mining
Q2: What are the most commonly used words in negative Chinese restaurant reviews?

Top 20 words Weights

Can you figure out why the customers were negative?


Questions # 3 – Topic Modeling via Text Mining
Q3: What are the topics when customers talk about Chinese restaurants in Las Vegas?

Can you figure out the meaning of each topic?


Questions # 3 – Topic Modeling via Text Mining
Q3: What are the topics when customers talk about Chinese restaurants in Las Vegas?

Shanghai, Sichuan,
and Taiwan food

Cantonese food

American-Chinese
food

Positive

Negative

Can you figure out the meaning of each topic?


Questions # 3 – Topic Modeling via Text Mining

Relatedness probability {0.01, 0.8, 0.001, 0.7, 0.5}

0.01

0.8

0.001

0.7

0.5
Questions # 3 – Topic Modeling via Text Mining

Relatedness probability {0.8, 0.01, 0.001, 0.7, 0.4}

0.8

0.01

0.001

0.7

0.4
Questions # 3 – Topic Modeling via Text Mining
1. Shanghai, Sichuan, and Taiwan food
2. Cantonese food
Use the relatedness probability to calculate similarity
3. American-Chinese food
between the two restaurants
4. Positive
5. Negative • Are the two restaurants similar in terms of cuisine?

Relatedness probability {0.01, 0.8, 0.001, 0.7, 0.5}

X X X

Relatedness probability {0.8, 0.01, 0.001, 0.7, 0.4}

0.008
+0.008
+0.00001
=0.01601
Questions # 3 – Topic Modeling via Text Mining
1. Shanghai, Sichuan, and Taiwan food
2. Cantonese food
Use the relatedness probability to calculate similarity
3. American-Chinese food
between the two restaurants
4. Positive
5. Negative
• Are the two restaurants similar in terms of customer satisfaction?

Relatedness probability {0.01, 0.8, 0.001, 0.7, 0.5}

X X

Relatedness probability {0.8, 0.01, 0.001, 0.7, 0.4}

0.49
+0.02
=0.51
Decision Making of Opening a Business
1. What is the average customer’s attitude toward Chinese restaurants in Las Vegas?
2. What are the most commonly used words in negative Chinese restaurant reviews?
3. What are the aspects when customers talk about Chinese restaurants in Las Vegas?
Is it a good idea to open a Chinese restaurant in Las Vegas? If so, how to do it?

Any implications for customer


relationship management?

• Cantonese food
• Shanghai, Sichuan, and Taiwan food
• American-Chinese food
• Positive
• Negative
Network Example - Facebook
A Simple Network

Node Strength of a link


B A

Link
Network Analysis – Organization Structure

• While traditionally management uses the


official chain of command to understand the
inner structure of an organization, it is
increasingly evident that the informal
network, capturing who really communicates
with whom, matters even more for the
success of a company.

• Accurate maps of this network can expose lack


of communication between key units, can
identify individuals who play an outsize role in
bringing different departments and products
together, and help higher management
diagnose diverse organizational issues.
Network Analysis – Organization Structure
• A Hungarian company that has three locations, one on
Budapest and two others outside of the city.

• list of a company’s employees, colored based on their


locations.
Network Analysis – Organization Structure
To apply network science, we need a network:
• Whom do you ask for advice when it comes to decisions
that impact the company, like, restructuring, advancement,
and so on?
Network Analysis – Organization Structure

• The bigger nodes represent


employees who has more
influence in the communication.

• Guess who they are?

• Do you see any problem in their


organizational communication?
Network Analysis – Consumer Engagement

• Smartone, a telecommunication
provider, releases a new service.

• There are a couple of users


adopting this new service.

• However, there are still many


users that don’t.

• How to identify and send Node: Smartone user


Link: Phone call between users
promotions to the potential users Link strength: #calls
who are non adopters?
Network Analysis – Social Influence on Your Health

Your friend running records

How will such records


affect your running
behavior?
Network Analysis – Social Influence on Your Health

Node: runner
Link: friendship

Running trace in the NYC central park


Network Analysis – Social Influence on Your Health

Influence

Which type of your friends make you run more?


• Active runner?
• Inactive runner?
Network Analysis – Social Influence on Your Health

Influence

How about gender?


• Female friends -> Female [you]
• Female friends-> Male [you]
• Male friends-> Female [you]
• Male friends-> Male [you]
Pipeline to do Text Mining and Network Analysis

1. What are the contexts/problems where


you can leverage text mining & Focus of ISOM 2010
network analysis?

2. Data collection

3. Data cleaning
Learn these techniques in data
4. Sentiment analysis analytics courses via Python!
Not easy but very rewarding!

5. Topic modeling

6. Network construction & analysis


Lecture Roadmap




Let me know if I can help!
Suggestions are always welcome!

Zhitao Yin
Office: LSK Room 4058
Email: zhitaoyin@ust.hk

You might also like