ELEC 6910Q Analytics and Systems for Social Media and Big Data Applications

Size: px
Start display at page:

Download "ELEC 6910Q Analytics and Systems for Social Media and Big Data Applications"

Transcription

1 ELEC 6910Q Analytics and Systems for Social Media and Big Data Applications Lecture 1 Intro. to Analytics, Big Data, and this Course Prof. James She james.she@ust.hk 1

2 About me Assistant Professor, ECE, HKUST Founding Director, HKUST-NIE Social Media Lab., ECE, HKUST. Associate Editor, ACM Trans. on Multimedia Comp., Comm. and Apps. (TOMM) Research interests: Social media, network, systems and anylitics with big data Multimedia Analytics through deep learning Wearable social/interactive/cyber-physical media systems Internet of Things infrastructure and applications 2

3 Course Overview Analytics Social Media and Network Analysis Big Data Analytics and Systems Useful Tools and Practical Skills 3

4 Outcomes of this lecture 1. What are Analytics? 2. What is Big Data? 3. About This Course 4. Social Media and Social Graph (if timing allows) Tag clouds using words from Chen, Hsinchun, Roger HL Chiang, and Veda C. Storey. "Business Intelligence and Analytics: From Big Data to Big Impact." MIS quarterly 36.4 (2012):

5 1. What are Analytics? 5

6 What Are Analytics Better Insight of Data Better Decisions/ Performances Discover new predictive or hidden patterns from data develop model, algorithm and method that is able to be verified better performances, automation, save costs, more revenue, creative solutions, etc. understand the data with existing and new statistical, mathematical, and scientific process etc. 6

7 Applications of Analytics - 1 Example: Recommendation or Personalization A social graph (SG): A C Friendship Content sharing (e.g., images) Info. or item recommendations B D 7

8 Applications of Analytics - 2 Example: Media and AI: Adjust Outdoor Ads for Optimal Responses Trained by faces and combintations with genders and ages ML Engine ML Engine 8

9 Applications of Analytics - 3 Example: Smart City and Life through AI 9

10 Analytics Platforms Varying by platforms, natures and purposes S/W (+ H/W) to provide tools and computational power Click analytics: interactive activities, e.g., likes/following/fans Descriptive: graphically report static or real-time data. Predictive: predict possible trends or future events 10

11 Descriptive Analytics Summary of Big Data in Real-time Collect data from different sources (social media, access log) Visualize the data for better decision 11

12 Predictive Analytics Prediction for Futures Utilizes statistical, modeling, data mining, machine learning, etc. Investigate or correlate recent and historical data Make prediction or discover new and predictable correlations 12

13 Advanced Analytics Connection Discovery Learning the user relationships through images 13

14 2. What & Where is Big Data? Why now? 14

15 What & Where is Big Data? 15

16 Big Data in Social Media Using shared images on social media, user s connections and preferences can be learnt Big Data Analytics Learn user connections! 16

17 Big Data in Enterprise Many open-source/free software for commercial applications computation storage analytics/ deep learning 17

18 Big Data in IoT BLE Beacons as IoT Infrastructure Mainly used for micro-positioning applications They can provide: 1. environmental sensor data 2. heat map of user activity 18

19 4Vs in Big Data Common but vague definition 19

20 Regarding Data Formats of Data Text Image Audio Video Emoji Type of Information Social relationship Sentiment Personal preference Activity Location Temperature & Humidity Heart rate Time/date User ID/address 20

21 What is Big Data? How Big is Big? MUCCCH larger / more complex dataset than traditional data that conventional processing S/W + H/W are inadequate. Developing analytics that are unmanagable or unknown before New challenges in processing, analysis, capture, search, sharing, storage, transfer, visualization, etc. 21

22 Why now? 22

23 Analytics and Systems for Social Media and Big Data Applications Social Media IoT Enterprises Applications Analytics Big Data Systems Computation Storage Delivery Big Data /Cloud Infrastructures 23

24 What is Big Data Analytics? Past Data mining 24

25 What is Big Data Analytics? Now and coming cloud-based computing 25

26 Big Data Systems Scalable Computation Massively Parallel Processing Distributed and Executed a command on multiple processors Coordinate distributed computing and storage resources 26

27 Emerging Infrastructures Big Data Analytics and Infrastructure as a Service (BDaaS) Providers Virtual hardware and software service for storage and analysis More cost-efficient and time-saving Cutting-edge technology are more accessible 27

28 Emerging Infrastructures Services provided by BDaaS Amazon provides different tools for handling big data computation storage analytics 28

29 Course Overview Some key info. 29

30 Course Overview Tutorial can help your programming and practical bits 30

31 Course Overview Reminders NO exams, but involves programming efforts and projects. But, we will help you through tutorials! Think twice! Avg. Grade Dist. ( ) Grade Percentage A+ 9.5% A 28.6% A- 28.6% B+ 19.0% B 14.3% > 50% 31

32 Project Papers Publications 32

33 Tools and Techniques That You Will Learn! community detection Stochastic gradient descent loss function in deep learning And much more 33

34 Opportunities After Taking ELEC 6910Q 34

35 Job Posting from MBDA Title: Data Scientists & Engineers Roles: Build models, frameworks, systems and products based on client's requirements and data Requirements: Proficiency in Python Proficiency in deep learning and related libraries Knowledge in NLP/computer vision/audio processing is a plus 35

36 Recall: Analytics and Systems for Social Media and Big Data Applications Social Media IoT Enterprises Applications Analytics Big Data Systems Computation Storage Delivery Big Data /Cloud Infrastructures 36

37 Announcements before moving on 1. Course Facebook page: search ELEC6910Q 2. If you decide to take this course, google Gephi and Python before the tutorial. 3. Check the location of tutorial, don t be late. 37

38 In-class Activity 1 (Group of 2-3) Give 2 Examples of Big Data and 1. Identify whether the data is from Social Media, Enterprise, or IoT 2. Identify the format of data and types of information 3. Identify 2 idea of possible knowledge or insight that can be learnt, analyzed or predicted from this Big Data. Submit your response on the In-class activity post on course Facebook Page 38

39 Sample Answer Instagram Uber Source Social media Enterprise Format of data Text, image timestamp, sensor data Types of information Insight User profile, user shared images, comment Social connection, user interest Prices, rating, GPS Route recommendation, trajectory 39

40 10 min break 40

41 What is Social Media? 41

42 Social Media Statistics 42

43 Social Media Multimedia Big Data 43

44 Social Media What s that? Why it matters? social media (content) has been generated since your 1 st social networking with others using Why it matters now? 44

45 Traditional Media The scale/style communications and networking 45

46 Social Media What s unique? 46

47 Social Media Not just devices/systems with media content 1. Publish by any user, then filter by users consumers == producers collaborative access, sharing, decision, etc. collective wisdom long tail 2. BIG amount of Data understanding about user profiles relationships/interactions & types locations 47

48 - End of Lecture 1-48