Introduction to Data Analytics Current Challenges Course Outline

Introduction to Data Analytics, Current Challenges. Course Outline Peter Fox Data Analytics – ITWS-4963/ITWS-6965 Week 1 a, January 21, 2014, SAGE 3101 1

Admin info (keep/ print this slide) • • • Class: ITWS-4963/ITWS 6965 Hours: 12: 00 pm-1: 50 pm Tuesday/ Friday Location: SAGE 3101 Instructor: Peter Fox Instructor contact: pfox@cs. rpi. edu, 518. 276. 4862 (do not leave a msg) Contact hours: Monday** 3: 00 -4: 00 pm (or by email appt) Contact location: Winslow 2120 (sometimes Lally 207 A announced by email) TA: Lakshmi Chenicheri chenil@rpi. edu Web site: http: //tw. rpi. edu/web/courses/Data. Analytics/2014 – Schedule, lectures, syllabus, reading, assignments, etc. 2

Contents • • Intro – about this course Learning objectives Outline of the course Definitions and why Analytics is more than Analysis • What skills are needed • What is expected 3

Truth in Advertising 4

Assessment and Assignments • Via written assignments with specific percentage of grade allocation provided with each assignment • Via individual oral presentations with specific percentage of grade allocation provided • Via presentations – depending on class size • Via participation in class (not to exceed 10% of total, start with 10% and lose % by not participating) • Late submission policy: first time with valid reason – no penalty, otherwise 20% of score deducted each late day. Talk to me EARLY if you are having schedule problems completing assignments 5

Assessment and Assignments • Reading assignments – Are given when needed to support key topics or to complete assignments – Will not be discussed in class unless there are questions • You will mostly perform individual work (i. e. group work is TBD) 6

Project options (examples) • • • Social networks Financial Social-economic, marketing Network/ security data Linked data 7

Objectives • Introduce students to relevant methods to recognize and apply quantitative algorithms, techniques and interpretation • To develop students' strategic thinking skills, combined with a solid technical foundation in data and model-driven decision-making. • Develop ability to apply critical and analytical methods to formulate and solve science, engineering, medical, and business problems • In groups, students will identify qualitative problems and apply content analytics • Students will examine real-world examples to place data-mining techniques in context, to develop data-analytic thinking, and to illustrate that proper application is as much an art as it is a science. • By the end of the course, students can effectively communicate analytic findings to non-specialists • [At the advanced level, evaluation focuses on decision making under uncertainty, learning how to build optimization models that incorporate random parameters: static stochastic optimization, two-stage optimization with recourse, chance-constrained optimization, and sequential decision making. ] 8

Learning Objectives • Through class lectures, practical sessions, written and oral presentation assignments and projects, students should: – Students to demonstrate knowledge of relevant analytic methods, and to recognize and apply quantitative algorithms, techniques and interpret results – Students to demonstrategic thinking skills, combined with a solid technical foundation in data and model-driven decision-making. – Students to develop ability to apply critical and analytical methods to formulate and solve science, engineering, medical, and business problems – Students will examine real-world examples to place data-mining techniques in context, to develop data-analytic thinking, and to illustrate that proper application is as much an art as it is a science. – Students must effectively communicate analytic findings to non-specialists. – [graduate level] Students must develop and demonstrate a working knowledge of decision making under uncertainty, be able to build optimization models that incorporate random parameters: static stochastic optimization, two-stage optimization with recourse, chance-constrained optimization, and sequential decision making. – ***TBD*** In groups, students will identify qualitative problems and apply content 9 analytics and present interpreted results

Undergraduates/ Grads • Graduate students are assessed at: – Higher level of demonstration – Additional questions or tasks in assignments • Undergraduates are welcome to complete these higher requirements to extra grade • Extra points for outstanding/ above and beyond are given** 10

Academic Integrity • Student-teacher relationships are built on trust. For example, students must trust that teachers have made appropriate decisions about the structure and content of the courses they teach, and teachers must trust that the assignments that students turn in are their own. Acts, which violate this trust, undermine the educational process. The Rensselaer Handbook of Student Rights and Responsibilities defines various forms of Academic Dishonesty and you should make yourself familiar with these. In this class, all assignments that are turned in for a grade must represent the student’s own work. In cases where help was received, or teamwork was allowed, a notation on the assignment should indicate your collaboration. • Submission of any assignment that is in violation of this policy will result in a penalty. If found in violation of the academic dishonesty policy, students may be subject to two types of penalties. The instructor administers an academic (grade) penalty of loss of grade for the work in violation, and the student may also enter the Institute judicial process and be subject to such additional sanctions as: warning, probation, suspension, expulsion, and alternative actions as defined in the current Handbook of Student Rights and Responsibilities. • If you have any question concerning this policy before submitting 11 an assignment, please ask for clarification.

Current Syllabus/Schedule • Web site: http: //tw. rpi. edu/web/courses/Data. Analytics/2 014 12

Questions so far? 13

Introductions • Who you are, background? • Why you are here? • What you expect to learn? 14

The nature of the challenge 15

16

17

Perspective • People make decisions every day and increasingly they are using computers to assist them. • Knowledge is power: – Or accurate/ reliable knowledge is actionable • Gaining knowledge and how to use that knowledge - from (often multiple ones) information and data sources 18

So what are we talking about? 19

Definitions (at least for this course) • Data - are encodings that represent the qualitative or quantitative attributes of a variable or set of variables. • Data (plural of "datum", which is seldom used) - are typically the results of measurements, computations, or observations and can be the basis of graphs, images of a set of variables. • Data - are often viewed as the lowest level of abstraction from which information and knowledge are derived*** 20

And then there is Big Data 21

~ Data for this course • Face. Book 100 - Facebook networks (from a date in Sept. 2005) for 100 colleges and universities. These files only include intraschool links. Anonymized. Well curated. Very good quality. Matlab. • Inter. Network Illinois – Telecommunication network traffic and telemetry for central Illinois. Well curated. Good quality. Unexplored. • Linked data – logd. tw. rpi. edu/datasets, e. g. EPA Facility Register System for each U. S. state. Linked. Quality unknown. RDF. 22

A view from IBM … • “Anyone who wants to learn something about data analytics should take a road trip. Myriad real-time decisions must be made based on analysis of static information as well as ever-changing conditions. Data about traffic, weather, road construction, fuel, time, current location and available funds are just a few of the factors. ” • This information and much more are needed to answer questions like: – If I skip this gas station, will I run out of gas before the next one? – Is it worth driving 50 miles out of the way to see the Corn Palace? How late will that side trip make us? – Can I make it to Billings, Mont. , by sunset or should I look for a place to stop? 23

Case Studies (warming up) • Sports Analytics – Moneyball (http: //www. imdb. com/title/tt 1210166/), Nate Silver (http: //en. wikipedia. org/wiki/Nate_Silver) • Google Analytics http: //www. marketingscoop. com/google-analyticscasestudy. htm • Marketing Analytics – products for pregnant (women) • Amazon Recommender – “If you liked, …” • http: //www. slideshare. net/lsakoda/case-studiesutilizing-real-time-data-analytics 24

Data… • Finding it • Using it: – Big data technologies: Hadoop (Map. Reduce) + Pig, HDFS, HIVE – see http: //projects. apache. org/indexes/category. html# database – No. SQL, Graph, Hbase, Cassandra, Mongo DB, Riak, Couch. DB – MPP Databases: Storm, Drill, Dremel • We’ll install/ use a few of these 25

Analysis • Software packages / environments: • • Gnu R Scipy Matlab Pig • Going from preliminary to initial analysis… • Parametric (assumes or asserts a probability distribution) and non-parametric statistics 26

What is "statistics"? • The term "statistics" has two common meanings, which we want to clearly separate: descriptive and inferential statistics. • But to understand the difference between descriptive and inferential statistics, we must first be clear on the difference between populations and samples. Courtesy Marshall Ma (and prior sources) 27

Populations and samples • A population is a set of well-defined objects – We must be able to say, for every object, if it is in the population or not – We must be able, in principle, to find every individual of the population A geographic example of a population is all pixels in a multi-spectral satellite image • A sample is a subset of a population – We must be able to say, for every object in the population, if it is in the sample or not – Sampling is the process of selecting a sample from a population Continuing the example, a sample from this population could be a set of pixels from known ground truth points Courtesy Marshall Ma (and prior sources) 28

What do we mean by "statistics"? • Two common uses of the word: • Descriptive statistics: numerical summaries of samples – i. e. , what was observed – Note the ‘sample’ may be exhaustive, i. e. , identical to the population • Inferential statistics: from samples to populations – i. e. , what could have been or will be observed in a larger population Example: Descriptive "The adjustments of 14 GPS control points for this orthorectification ranged from 3. 63 to 8. 36 m with an arithmetic mean of 5. 145 m" Inferential "The mean adjustment for any set of GPS points taken under specified conditions and used for orthorectification is no less than 4. 3 and no more than 6. 1 m; this statement has a 5% probability of being wrong" Courtesy Marshall Ma (and prior sources) 29

Patterns and Relationships • Stepping from elementary/ distribution analysis to algorithmic-based analysis • Often: data mining: classification, clustering, rules; machine learning; support vector machine, non-parametric models • Outcome: model and an evaluation of its fitness for purpose 30

Prediction • • Choosing applicable models Combining models Confidence levels Multi-variate • Future, event, pattern • Past event, relation • Etc. 31

Prescription • Decisions and Effects: What should you do and why? “Business Rules” • Benefit or mitigate or adapt? [Personal e. g. ] • Builds on Prediction, often involves scenarios and post-analysis 32

Summary • We’ll work our way through the stages of analytics • We’ll use current data infrastructures for analytics to give you practical experience • We’ll cover algorithms, models, and software to use them • Aim: Tuesday lecture, Friday hands-on and interactions • Course will be adapted as we go 33

Just for fun… 34

Traversal for new patterns 35

Smart visual exploration 36

Skills needed • Database or data structures? • Literacy with computers and applications that can handle the data we will use • Ability to access internet, servers and retrieve/ acquire data • Presentation of proposal projects and assignment results 37

Tentative assignment structure (no exam) • Assignment 1: Review of a DA Case Study. Due ~ week 2 (Friday). 10% (written/ discuss; individual); • Assignment 2: Datasets and data infrastructures – lab assignment. Due ~ week 3. 10% (lab; individual); • Assignment 3: Preliminary and Statistical Analysis. Due ~ week 4. 15% (15% written and 0% oral; individual); • Assignment 4: Pattern, trend, relations: model development and evaluation. Due ~ week 5. 15% (10% written and 5% oral; individual); • Assignment 5: Term project proposal. Due ~ week 6. 5% (0% written and 5% oral; individual); • Assignment 6: Predictive and Prescriptive Analytics. Due ~ week 8. 15% (15% written and 5% oral; individual); 38 • Term project. Due ~ week 13. 30% (25% written, 5% oral; individual).

What is expected • Attend class, complete assignments, participate • Ask questions, offer answers in class • Work individually (and in a group TBD) on assignments • Work constructively in class sessions • Next class is Jan. 24 – Getting some applications setup/ running 39

Reading/ watching • Sports Analytics – Moneyball (http: //www. imdb. com/title/tt 1210166/), • Nate Silver (http: //en. wikipedia. org/wiki/Nate_Silver) • Google Analytics http: //www. marketingscoop. com/google-analyticscasestudy. htm • http: //www. slideshare. net/lsakoda/case-studiesutilizing-real-time-data-analytics • http: //www. marketquotient. com/case-studies. html • http: //www. ibm. com/analytics/us/en/case-studies/ 40

Assignment 1 • Choose a DA case study from a) readings, or b) your choice (must be approved by me) • Read it and provide a short written review/ critique (business case, area of application, approach/ methods, tools used, results, actions, benefits). Hand in the written report. • Be prepared to discuss it on class on Friday 31 st. • Details on the Web site (under Reading/Assignments; Week 1) 41