Praise for Mining the Social Web “Mining the Social Web is a must-read as data is distributed at a dizzying pace. A great primer for API jockeys, social media junkies, and data scientists alike, [Matthew] Russell deftly distills the prodigious opportunity in mining social media data.” — Nick Ducoff, CEO of Infochimps, Inc. “This is an essential guide to tapping the newgeneration of online data sources. Russell has done a great job creating an accessible manual for anyone working with social informa- tion on the web, covering both how to access it and simple methods for extracting surprising insights from all that raw data.” — Pete Warden, Founder of OpenHeatMap.com “Mining the Social Web is nowmy go-to book for any project that involves analyzing social data. It contains a multitude of useful examples and is highly recommended for any data mining project you’re considering. Great for beginners and advanced readers alike.” — Abe Music, Principal, Zaffra “This book is clearly a labor of love for the author. He has deftly woven together the use of classic text and graph mining libraries with current social media applications. Examples are concrete and concise while providing useful insights that facilitate future development and exploration by the reader. This text is a great primer for those just beginning their forays into extracting understanding from social networks, and also for advanced researchers needing access to the latest social media APIs.” — Chris Augeri, Senior Research Fellow, University of Nebraska “This is a phenomenal book for anyone wanting to get started mining social data. It is well- researched and provides plenty of examples to get one going from the very first chapter. It is also very easy to followand a real pleasure to read. This book is my first recommendation for anyone interested in the mining, analysis, and visualization of data from the social web.” — Jeffrey Humphries, PhD; Computer Scientist Mining_praise_page Page i Wednesday, January 12, 2011 10:28 AM “Few things will impact us the way automated understanding of human communication by software will in the coming years. This subject is broad and deep. It has been the subject of thousands of papers and hundreds of dissertations. What Matthewhas pulled together is something that has really been missing: an applied introduction to a diverse and deep set of technologies and topics that make the knowledge buried in human communication inside the social web accessible. It is the work of a powerful technologist—someone who can equip capable programmers with new tools that are truly valuable. Read this book. It will open up doors to where software is going in the next decade.” — Tim Estes, Founder and CEO, Digital Reasoning “Mining the Social Web is a great resource on how to get the most out of the Twitter API.” — Raffi Krikorian, Platform Services group, Twitter “Matthewcovers an interesting and eclectic group of data sources, analysis techniques, data management tools, and visualizations that provide a thorough survey of the latest thinking on howto gain insight from the social web.His examples are vivid and serve as great starting points for further exploration. Matthewclearly cares that the reader under- stands the material; the book is chock full of timely, knowing, and truly helpful hints and advice. Mining the Social Web has me excited to dive further into this rich area of analysis.” — Roger Magoulas, Director of Market Research, O’Reilly Media Mining_praise_page Page ii Wednesday, January 12, 2011 10:28 AM Mining the Social Web Mining the Social Web Matthew A. Russell Beijing • Cambridge • Farnham • Köln • Sebastopol • Tokyo Mining the Social Web by Matthew A. Russell Copyright © 2011 Matthew Russell. All rights reserved. Printed in the United States of America. Published by O’Reilly Media, Inc., 1005 Gravenstein Highway North, Sebastopol, CA 95472. O’Reilly books may be purchased for educational, business, or sales promotional use. Online editions are also available for most titles (http://my.safaribooksonline.com). For more information, contact our corporate/institutional sales department: (800) 998-9938 or corporate@oreilly.com. Editor: Mike Loukides Production Editor: Adam Zaremba Copyeditor: Rachel Head Proofreader: Marlowe Shaeffer Indexer: Ellen Troutman Zaig Cover Designer: Karen Montgomery Interior Designer: David Futato Illustrator: Robert Romano Printing History: January 2011: First Edition. Nutshell Handbook, the Nutshell Handbook logo, and the O’Reilly logo are registered trademarks of O’Reilly Media, Inc. Mining the Social Web, the image of a groundhog, and related trade dress are trade- marks of O’Reilly Media, Inc. Many of the designations used by manufacturers and sellers to distinguish their products are claimed as trademarks. Where those designations appear in this book, and O’Reilly Media, Inc. was aware of a trademark claim, the designations have been printed in caps or initial caps. While every precaution has been taken in the preparation of this book, the publisher and author assume no responsibility for errors or omissions, or for damages resulting from the use of the information con- tained herein. TM This book uses RepKover™, a durable and flexible lay-flat binding. ISBN: 978-1-449-38834-8 [M] 1294936576 To those seeking knowledge and wisdom: Use wisdom and understanding to establish your home; Let good sense fill the rooms with priceless treasures. Wisdom brings strength, and knowledge gives power. Battles are won by listening to advice and making a lot of plans. May you find knowledge and wisdom. Table of Contents Preface . ................................................................... xiii 1. Introduction: Hacking on Twitter Data . ..................................... 1 Installing Python Development Tools 1 Collecting and Manipulating Twitter Data 3 Tinkering with Twitter’s API 4 Frequency Analysis and Lexical Diversity 7 Visualizing Tweet Graphs 14 Synthesis: Visualizing Retweets with Protovis 15 Closing Remarks 17 2. Microformats: Semantic Markup and Common Sense Collide . . . . . . . . . . . . . . . . . . 19 XFN and Friends 19 Exploring Social Connections with XFN 22 A Breadth-First Crawl of XFN Data 23 Geocoordinates: A Common Thread for Just About Anything 30 Wikipedia Articles + Google Maps = Road Trip? 30 Slicing and Dicing Recipes (for the Health of It) 35 Collecting Restaurant Reviews 37 Summary 40 3. Mailboxes: Oldies but Goodies . .......................................... 41 mbox: The Quick and Dirty on Unix Mailboxes 42 mbox + CouchDB = Relaxed Email Analysis 48 Bulk Loading Documents into CouchDB 51 Sensible Sorting 52 Map/Reduce-Inspired Frequency Analysis 55 Sorting Documents by Value 61 couchdb-lucene: Full-Text Indexing and More 63 Threading Together Conversations 67 Look Who’s Talking 73 ix Visualizing Mail “Events” with SIMILE Timeline 77 Analyzing Your Own Mail Data 80 The Graph Your (Gmail) Inbox Chrome Extension 81 Closing Remarks 82 4. Twitter: Friends, Followers, and Setwise Operations . . . . . . . . . . . . . . . . . . . . . . . . . 83 RESTful and OAuth-Cladded APIs 84 No, You Can’t Have My Password 85 A Lean, Mean Data-Collecting Machine 88 A Very Brief Refactor Interlude 91 Redis: A Data Structures Server 92 Elementary Set Operations 94 Souping Up the Machine with Basic Friend/Follower Metrics 96 Calculating Similarity by Computing Common Friends and Followers 102 Measuring Influence 103 Constructing Friendship Graphs 108 Clique Detection and Analysis 110 The Infochimps “Strong Links” API 114 Interactive 3D Graph Visualization 116 Summary 117 5. Twitter: The Tweet, the Whole Tweet, and Nothing but the Tweet . . . . . . . . . . . . 119 Pen : Sword :: Tweet : Machine Gun (?!?) 119 Analyzing Tweets (One Entity at a Time) 122 Tapping (Tim’s) Tweets 125 Who Does Tim Retweet Most Often? 138 What’s Tim’s Influence? 141 How Many of Tim’s Tweets Contain Hashtags? 144 Juxtaposing Latent Social Networks (or #JustinBieber Versus #TeaParty) 147 What Entities Co-Occur Most Often with #JustinBieber and #TeaParty Tweets? 148 On Average, Do #JustinBieber or #TeaParty Tweets Have More Hashtags? 153 Which Gets Retweeted More Often: #JustinBieber or #TeaParty? 154 How Much Overlap Exists Between the Entities of #TeaParty and #JustinBieber Tweets? 156 Visualizing Tons of Tweets 158 Visualizing Tweets with Tricked-Out Tag Clouds 158 Visualizing Community Structures in Twitter Search Results 162 Closing Remarks 166 6. LinkedIn: Clustering Your Professional Network for Fun (and Profit?) . . . . . . . . . . 167 Motivation for Clustering 168 x | Table of Contents Clustering Contacts by Job Title 172 Standardizing and Counting Job Titles 172 Common Similarity Metrics for Clustering 174 A Greedy Approach to Clustering 177 Hierarchical and k-Means Clustering 185 Fetching Extended Profile Information 188 Geographically Clustering Your Network 193 Mapping Your Professional Network with Google Earth 193 Mapping Your Professional Network with Dorling Cartograms 198 Closing Remarks 198 7. Google Buzz: TF-IDF, Cosine Similarity, and Collocations . . . . . . . . . . . . . . . . . . . . . 201 Buzz = Twitter + Blogs (???) 202 Data Hacking with NLTK 205 Text Mining Fundamentals 209 A Whiz-Bang Introduction to TF-IDF 209 Querying Buzz Data with TF-IDF 215 Finding Similar Documents 216 The Theory Behind Vector Space Models and Cosine Similarity 217 Clustering Posts with Cosine Similarity 219 Visualizing Similarity with Graph Visualizations 222 Buzzing on Bigrams 224 How the Collocation Sausage Is Made: Contingency Tables and Scoring Functions 228 Tapping into Your Gmail 231 Accessing Gmail with OAuth 232 Fetching and Parsing Email Messages 233 Before You Go Off and Try to Build a Search Engine… 235 Closing Remarks 237 8. Blogs et al.: Natural Language Processing (and Beyond) . . . . . . . . . . . . . . . . . . . . . 239 NLP: A Pareto-Like Introduction 239 Syntax and Semantics 240 A Brief Thought Exercise 241 A Typical NLP Pipeline with NLTK 242 Sentence Detection in Blogs with NLTK 245 Summarizing Documents 250 Analysis of Luhn’s Summarization Algorithm 256 Entity-Centric Analysis: A Deeper Understanding of the Data 258 Quality of Analytics 267 Closing Remarks 269 Table of Contents | xi D o wnload from Wow! eBook
9. Facebook: The All-in-One Wonder . ...................................... 271 Tapping into Your Social Network Data 272 From Zero to Access Token in Under 10 Minutes 272 Facebook’s Query APIs 278 Visualizing Facebook Data 289 Visualizing Your Entire Social Network 289 Visualizing Mutual Friendships Within Groups 301 Where Have My Friends All Gone? (A Data-Driven Game) 304 Visualizing Wall Data As a (Rotating) Tag Cloud 309 Closing Remarks 311 10. The Semantic Web: A Cocktail Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313 An Evolutionary Revolution? 313 Man Cannot Live on Facts Alone 315 Open-World Versus Closed-World Assumptions 315 Inferencing About an Open World with FuXi 316 Hope 319 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321 xii | Table of Contents Preface The Web is more a social creation than a technical one. I designed it for a social effect—to help people work together—and not as a technical toy. The ultimate goal of the Web is to support and improve our weblike exis- tence in the world. We clump into families, associations, and companies. We develop trust across the miles and distrust around the corner. —Tim Berners-Lee, Weaving the Web (Harper) To Read This Book? If you have a basic programming background and are interested in insight surrounding the opportunities that arise from mining and analyzing data from the social web, you’ve come to the right place. We’ll begin getting our hands dirty after just a few more pages of frontmatter. I’ll be forthright, however, and say upfront that one of the chief com- plaints you’re likely to have about this book is that all of the chapters are far too short. Unfortunately, that’s always the case when trying to capture a space that’s evolving daily and is so rich and abundant with opportunities. That said, I’m a fan of the “80-20 rule”, and I sincerely believe that this book is a reasonable attempt at presenting the most interesting 20 percent of the space that you’d want to explore with 80 percent of your available time. This book is short, but it does cover a lot of ground. Generally speaking, there’s a little more breadth than depth, although where the situation lends itself and the subject matter is complex enough to warrant a more detailed discussion, there are a few deep dives into interesting mining and analysis techniques. The book was written so that you could have the option of either reading it from cover to cover to get a broad primer on working with social web data, or pick and choose chapters that are of particular interest to you. In other words, each chapter is designed to be bite-sized and fairly standalone, but special care was taken to introduce material in a particular order so that the book as a whole is an enjoyable read. xiii Social networking websites such as Facebook, Twitter, and LinkedIn have transitioned from fad to mainstream to global phenomena over the last few years. In the first quarter of 2010, the popular social networking site Facebook surpassed Google for the most page visits,* confirming a definite shift in how people are spending their time online. Asserting that this event indicates that the Web has now become more a social milieu than a tool for research and information might be somewhat indefensible; however, this data point undeniably indicates that social networking websites are satisfying some very basic human desires on a massive scale in ways that search engines were never designed to fulfill. Social networks really are changing the way we live our lives on and off the Web,† and they are enabling technology to bring out the best (and sometimes the worst) in us. The explosion of social networks is just one of the ways that the gap between the real world and cyberspace is continuing to narrow. Generally speaking, each chapter of this book interlaces slivers of the social web along with data mining, analysis, and visualization techniques to answer the following kinds of questions: • Who knows whom, and what friends do they have in common? • How frequently are certain people communicating with one another? • How symmetrical is the communication between people? • Who are the quietest/chattiest people in a network? • Who are the most influential/popular people in a network? • What are people chatting about (and is it interesting)? The answers to these types of questions generally connect two or more people together and point back to a context indicating why the connection exists. The work involved in answering these kinds of questions is only the beginning of more complex analytic processes, but you have to start somewhere, and the low-hanging fruit is surprisingly easy to grasp, thanks to well-engineered social networking APIs and open source toolkits. Loosely speaking, this book treats the social web‡ as a graph of people, activities, events, concepts, etc. Industry leaders such as Google and Facebook have begun to increasingly push graph-centric terminology rather than web-centric terminology as they simulta- neously promote graph-based APIs. In fact, Tim Berners-Lee has suggested that perhaps he should have used the term Giant Global Graph (GGG) instead of World Wide Web (WWW), because the terms “web” and “graph” can be so freely interchanged in the context of defining a topology for the Internet. Whether the fullness of Tim Berners- * See the opening paragraph of Chapter 9. † Mark Zuckerberg, the creator of Facebook, was named Person of the Year for 2010 by Time magazine (http: //www.time.com/time/specials/packages/article/0,28804,2036683_2037183_2037185,00.html) ‡ See http://journal.planetwork.net/article.php?lab=reed0704 for another perspective on the social web that focuses on digital identities. xiv | Preface Lee’s original vision will ever be realized remains to be seen, but the Web as we know it is getting richer and richer with social data all the time. When we look back years from now, it may well seem obvious that the second- and third-level effects created by an inherently social web were necessary enablers for the realization of a truly semantic web. The gap between the two seems to be closing. Or Not to Read This Book? Activities such as building your own natural language processor from scratch, venturing far beyond the typical usage of visualization libraries, and constructing just about any- thing state-of-the-art are not within the scope of this book. You’ll be really disappointed if you purchase this book because you want to do one of those things. However, just because it’s not realistic or our goal to capture the holy grail of text analytics or record matching in a mere few hundred pages doesn’t mean that this book won’t enable you to attain reasonable solutions to hard problems, apply those solutions to the social web as a domain, and have a lot of fun in the process. It also doesn’t mean that taking a very active interest in these fascinating research areas wouldn’t potentially be a great idea for you to consider. A short book like this one can’t do much beyond whetting your appetite and giving you enough insight to go out and start making a difference some- where with your newly found passion for data hacking. Maybe it’s obvious in this day and age, but another important item of note is that this book generally assumes that you’re connected to the Internet. This wouldn’t be a great book to take on vacation with you to a remote location, because it contains many references that have been hyperlinked, and all of the code examples are hyperlinked directly to GitHub, a very social Git repository that will always reflect the most up-to- date example code available. The hope is that social coding will enhance collaboration between like-minded folks such as ourselves who want to work together to extend the examples and hack away at interesting problems. Hopefully, you’ll fork, extend, and improve the source—and maybe even make some new friends along the way. Readily accessible sources of online information such as API docs are also liberally hyperlinked, and it is assumed that you’d rather look them up online than rely on inevitably stale copies in this printed book. The official GitHub repository that maintains the latest and greatest bug-fixed source code for this book is http://github.com/ptwobrussell/ Mining-the-Social-Web. The official Twitter account for this book is @SocialWebMining. This book is also not recommended if you need a reference that gets you up to speed on distributed computing platforms such as sharded MySQL clusters or NoSQL tech- nologies such as Hadoop or Cassandra. We do use some less-than-conventional storage technologies such as CouchDB and Redis, but always within the context of running on Preface | xv a single machine, and because they work well for the problem at hand. However, it really isn’t that much of a stretch to port the examples into distributed technologies if you possess sufficient motivation and need the horizontal scalability. A strong recom- mendation is that you master the fundamentals and prove out your thesis in a slightly less complex environment first before migrating to an inherently more complex dis- tributed system—and then be ready to make major adjustments to your algorithms to make them performant once data access is no longer local. A good option to investigate if you want to go this route is Dumbo. Stay tuned to this book’s Twitter account (@SocialWebMining) for extended examples that involve Dumbo. This book provides no advice whatsoever about the legal ramifications of what you may decide to do with the data that’s made available to you from social networking sites, although it does sincerely attempt to comply with the letter and spirit of the terms governing the particular sites that are mentioned. It may seem unfortunate that many of the most popular social networking sites have licensing terms that prohibit the use of their data outside of their platforms, but at the moment, it’s par for the course. Most social networking sites are like walled gardens, but from their standpoint (and the standpoint of their investors) a lot of the value these companies offer currently relies on controlling the platforms and protecting the privacy of their users; it’s a tough bal- ance to maintain and probably won’t be all sorted out anytime soon. A final and much lesser caveat is that this book does slightly favor a *nix environ- ment,§ in that there are a select few visualizations that may give Windows users trouble. Whenever this is known to be a problem, however, advice is given on reasonable al- ternatives or workarounds, such as firing up a VirtualBox to run the example in a Linux environment. Fortunately, this doesn’t come up often, and the few times it does you can safely ignore those sections and move on without any substantive loss of reading enjoyment. Tools and Prerequisites The only real prerequisites for this book are that you need to be motivated enough to learn some Python and have the desire to get your hands (really) dirty with social data. None of the techniques or examples in this book require significant background knowl- edge of data analysis, high performance computing, distributed systems, machine learning, or anything else in particular. Some examples involve constructs you may not have used before, such as thread pools, but don’t fret—we’re programming in Python. Python’s intuitive syntax, amazing ecosystem of packages for data manipulation, and core data structures that are practically JSON make it an excellent teaching tool that’s powerful yet also very easy to get up and running. On other occasions we use some packages that do pretty advanced things, such as processing natural language, but we’ll § *nix is a term used to refer to a Linux/Unix environment, which is basically synonymous with non-Windows at this point in time. xvi | Preface D o wnload from Wow! eBook approach these from the standpoint of using the technology as an application pro- grammer. Given the high likelihood that very similar bindings exist for other program- ming languages, it should be a fairly rote exercise to port the code examples should you so desire. (Hopefully, that’s exactly the kind of thing that will happen on GitHub!) Beyond the previous explanation, this book makes no attempt to justify the selection of Python or apologize for using it, because it’s a very suitable tool for the job. If you’re new to programming or have never seen Python syntax, skimming ahead a few pages should hopefully be all the confirmation that you need. Excellent documentation is available online, and the official Python tutorial is a good place to start if you’re looking for a solid introduction. This book attempts to introduce a broad array of useful visualizations across a variety of visualization tools and toolkits, ranging from consumer staples like spreadsheets to industry staples like Graphviz, to bleeding-edge HTML5 technologies such as Proto- vis. A reasonable attempt has been made to introduce a couple of new visualizations in each chapter, but in a way that follows naturally and makes sense. You’ll need to be comfortable with the idea of building lightweight prototypes from these tools. That said, most of the visualizations in this book are little more than small mutations on out- of-the-box examples or projects that minimally exercise the APIs, so as long as you’re willing to learn, you should be in good shape. Conventions Used in This Book The following typographical conventions are used in this book: Italic Indicates new terms, URLs, email addresses, filenames, and file extensions. Constant width Indicates program listings, and is used within paragraphs to refer to program elements such as variable or function names, databases, data types, environment variables, statements, and keywords. Constant width bold Shows commands or other text that should be typed literally by the user. Also occasionally used for emphasis in code listings. Constant width italic Shows text that should be replaced with user-supplied values or values determined by context. This icon signifies a tip, suggestion, or general note. Preface | xvii This icon indicates a warning or caution. Using Code Examples Most of the numbered examples in the following chapters are available for download at GitHub at https://github.com/ptwobrussell/Mining-the-Social-Web—the official code repository for this book. You are encouraged to monitor this repository for the latest bug-fixed code as well as extended examples by the author and the rest of the social coding community. This book is here to help you get your job done. In general, you may use the code in this book in your programs and documentation. You do not need to contact us for permission unless you’re reproducing a significant portion of the code. For example, writing a program that uses several chunks of code from this book does not require permission. Selling or distributing a CD-ROM of examples from O’Reilly books does require permission. Answering a question by citing this book and quoting example code does not require permission. Incorporating a significant amount of example code from this book into your product’s documentation does require permission. We appreciate, but do not require, attribution. An attribution usually includes the title, author, publisher, and ISBN. For example: “Mining the Social Web by Matthew A. Russell. Copyright 2011 Matthew Russell, 978-1-449-38834-8.” If you feel your use of code examples falls outside fair use or the permission given above, feel free to contact us at permissions@oreilly.com. Safari® Books Online Safari Books Online is an on-demand digital library that lets you easily search over 7,500 technology and creative reference books and videos to find the answers you need quickly. With a subscription, you can read any page and watch any video from our library online. Read books on your cell phone and mobile devices. Access new titles before they are available for print, and get exclusive access to manuscripts in development and post feedback for the authors. Copy and paste code samples, organize your favorites, down- load chapters, bookmark key sections, create notes, print out pages, and benefit from tons of other time-saving features. O’Reilly Media has uploaded this book to the Safari Books Online service. To have full digital access to this book and others on similar topics from O’Reilly and other pub- lishers, sign up for free at http://my.safaribooksonline.com. xviii | Preface How to Contact Us Please address comments and questions concerning this book to the publisher: O’Reilly Media, Inc. 1005 Gravenstein Highway North Sebastopol, CA 95472 800-998-9938 (in the United States or Canada) 707-829-0515 (international or local) 707-829-0104 (fax) We have a web page for this book, where we list errata, examples, and any additional information. You can access this page at: http://oreilly.com/catalog/9781449388348/ Readers can request general help from the author and publisher through GetSatisfac- tion at: http://getsatisfaction.com/oreilly Readers may also file tickets for the sample code—as well as anything else in the book— through GitHub’s issue tracker at: http://github.com/ptwobrussell/Mining-the-Social-Web/issues To comment or ask technical questions about this book, send email to: bookquestions@oreilly.com For more information about our books, conferences, Resource Centers, and the O’Reilly Network, see our website at: http://www.oreilly.com Acknowledgments To say the least, writing a technical book takes a ridiculous amount of sacrifice. On the home front, I gave up more time with my wife, Baseeret, and daughter, Lindsay Belle, than I’m proud to admit. Thanks most of all to both of you for loving me in spite of my ambitions to somehow take over the world one day. (It’s just a phase, and I’m really trying to grow out of it—honest.) I sincerely believe that the sum of your decisions gets you to where you are in life (especially professional life), but nobody could ever complete the journey alone, and it’s an honor give credit where credit is due. I am truly blessed to have been in the company of some of the brightest people in the world while working on this book, including a technical editor as smart as Mike Loukides, a production staff as talented as the folks at O’Reilly, and an overwhelming battery of eager reviewers as amazing as everyone who helped me to complete this book. I especially want to thank Abe Music, Preface | xix Pete Warden, Tantek Celik, J. Chris Anderson, Salvatore Sanfilippo, Robert Newson, DJ Patil, Chimezie Ogbuji, Tim Golden, Brian Curtin, Raffi Krikorian, Jeff Hammer- bacher, Nick Ducoff, and Cameron Marlowe for reviewing material or making partic- ularly helpful comments that absolutely shaped its outcome for the best. I’d also like to thank Tim O’Reilly for graciously allowing me to put some of his Twitter and Google Buzz data under the microscope in Chapters 4, 5, and 7; it definitely made those chap- ters much more interesting to read than they otherwise would have been. It would be impossible to recount all of the other folks who have directly or indirectly shaped my life or the outcome of this book. Finally, thanks to you for giving this book a chance. If you’re reading this, you’re at least thinking about picking up a copy. If you do, you’re probably going to find some- thing wrong with it despite my best efforts; however, I really do believe that, in spite of the few inevitable glitches, you’ll find it an enjoyable way to spend a few evenings/ weekends and you’ll manage to learn a few things somewhere along the line. xx | Preface CHAPTER 1 Introduction: Hacking on Twitter Data Although we could get started with an extended discussion of specific social networking APIs, schemaless design, or many other things, let’s instead dive right into some intro- ductory examples that illustrate how simple it can be to collect and analyze some social web data. This chapter is a drive-by tutorial that aims to motivate you and get you thinking about some of the issues that the rest of the book revisits in greater detail. We’ll start off by getting our development environment ready and then quickly move on to collecting and analyzing some Twitter data. Installing Python Development Tools The example code in this book is written in Python, so if you already have a recent version of Python and easy_install on your system, you obviously know your way around and should probably skip the remainder of this section. If you don’t already have Python installed, the bad news is that you’re probably not already a Python hacker. But don’t worry, because you will be soon; Python has a way of doing that to people because it is easy to pick up and learn as you go along. Users of all platforms can find instructions for downloading and installing Python at http://www.python.org/down load/, but it is highly recommended that Windows users install ActivePython, which automatically adds Python to your path at the Windows Command Prompt (henceforth referred to as a “terminal”) and comes with easy_install, which we’ll discuss in just a moment. The examples in this book were authored in and tested against the latest Python 2.7 branch, but they should also work fine with other relatively up-to-date versions of Python. At the time this book was written, Python Version 2 is still the status quo in the Python community, and it is recommended that you stick with it unless you are confident that all of the dependencies you’ll need have been ported to Version 3, and you are willing to debug any idiosyncrasies involved in the switch. Once Python is installed, you should be able to type python in a terminal to spawn an interpreter. Try following along with Example 1-1. 1 Example 1-1. Your very first Python interpreter session >>> print "Hello World" Hello World >>> #this is a comment ... >>> for i in range(0,10): # a loop ... print i, # the comma suppresses line breaks ... 0 1 2 3 4 5 6 7 8 9 >>> numbers = [ i for i in range(0,10) ] # a list comprehension >>> print numbers [0, 1, 2, 3, 4, 5, 6, 7, 8, 9] >>> if 10 in numbers: # conditional logic ... print True ... else: ... print False ... False One other tool you’ll want to have on hand is easy_install,* which is similar to a package manager on Linux systems; it allows you to effortlessly install Python packages instead of downloading, building, and installing them from source. You can download the latest version of easy_install from http://pypi.python.org/pypi/setuptools, where there are specific instructions for each platform. Generally speaking, *nix users will want to sudo easy_install so that modules are written to Python’s global installation directories. It is assumed that Windows users have taken the advice to use ActivePy- thon, which automatically includes easy_install as part of its installation. Windows users might also benefit from reviewing the blog post “Installing easy_install…could be easier”, which discusses some com- mon problems related to compiling C code that you may encounter when running easy_install. Once you have properly configured easy_install, you should be able to run the fol- lowing command to install NetworkX—a package we’ll use throughout the book for building and analyzing graphs—and observe similar output: $ easy_install networkx Searching for networkx ...truncated output... Finished processing dependencies for networkx * Although the examples in this book use the well-known easy_install, the Python community has slowly been gravitating toward pip, another build tool you should be aware of and that generally “just works” with any package that can be easy_install’d. 2 | Chapter 1: Introduction: Hacking on Twitter Data With NetworkX installed, you might think that you could just import it from the in- terpreter and get right to work, but occasionally some packages might surprise you. For example, suppose this were to happen: >>> import networkx Traceback (most recent call last): ... truncated output ... ImportError: No module named numpy Whenever an ImportError happens, it means there’s a missing package. In this illus- tration, the module we installed, networkx, has an unsatisfied dependency called numpy, a highly optimized collection of tools for scientific computing. Usually, another invocation of easy_install fixes the problem, and this situation is no different. Just close your interpreter and install the dependency by typing easy_install numpy in the terminal: $ easy_install numpy Searching for numpy ...truncated output... Finished processing dependencies for numpy Now that numpy is installed, you should be able to open up a new interpreter, import networkx, and use it to build up graphs. Example 1-2 demonstrates. Example 1-2. Using NetworkX to create a graph of nodes and edges >>> import networkx >>> g=networkx.Graph() >>> g.add_edge(1,2) >>> g.add_node("spam") >>> print g.nodes() [1, 2, 'spam'] >>> print g.edges() [(1, 2)] At this point, you have some of your core Python development tools installed and are ready to move on to some more interesting tasks. If most of the content in this section has been a learning experience for you, it would be worthwhile to review the official Python tutorial online before proceeding further. Collecting and Manipulating Twitter Data In the extremely unlikely event that you don’t know much about Twitter yet, it’s a real- time, highly social microblogging service that allows you to post short messages of 140 characters or less; these messages are called tweets. Unlike social networks like Face- book and LinkedIn, where a connection is bidirectional, Twitter has an asymmetric network infrastructure of “friends” and “followers.” Assuming you have a Twitter Collecting and Manipulating Twitter Data | 3 account, your friends are the accounts that you are following and your followers are the accounts that are following you. While you can choose to follow all of the users who are following you, this generally doesn’t happen because you only want your Home Timeline† to include tweets from accounts whose content you find interesting. Twitter is an important phenomenon from the standpoint of its incredibly high number of users, as well as its use as a marketing device and emerging use as a transport layer for third- party messaging services. It offers an extensive collection of APIs, and although you can use a lot of the APIs without registering, it’s much more interesting to build up and mine your own network. Take a moment to review Twitter’s liberal terms of service, API documentation, and API rules, which allow you to do just about anything you could reasonably expect to do with Twitter data before doing any heavy-duty devel- opment. The rest of this book assumes that you have a Twitter account and enough friends/followers that you have data to mine. The official Twitter account for this book is @SocialWebMining. Tinkering with Twitter’s API A minimal wrapper around Twitter’s web API is available through a package called twitter (http://github.com/sixohsix/twitter) can be installed with easy_install per the norm: $ easy_install twitter Searching for twitter ...truncated output... Finished processing dependencies for twitter The package also includes a handy command-line utility and IRC bot, so after installing the module you should be able to simply type twitter in a shell to get a usage screen about how to use the command-line utility. However, we’ll focus on working within the interactive Python interpreter. We’ll work though some examples, but note that you can always skim the documentation by running pydoc from the terminal. *nix users can simply type pydoc twitter.Twitter to view the documentation on the Twitter class, while Windows users need to type python -mpydoc twitter.Twitter. If you find yourself reviewing the documentation for certain modules often, you can elect to pass the -w option to pydoc and write out an HTML page that you can save and bookmark in your browser. It’s also worth knowing that running pydoc on a module or class brings up the inline documentation in the same way that running the help() command in the inter- preter would. Try typing help(twitter.Twitter) in the interpreter to see for yourself. † http://support.twitter.com/entries/164083-what-is-a-timeline 4 | Chapter 1: Introduction: Hacking on Twitter Data D o wnload from Wow! eBook Without further ado, let’s find out what people are talking about by inspecting the trends available to us through Twitter’s search API. Let’s fire up the interpreter and initiate a search. Try following along with Example 1-3, and use the help() function as needed to try to answer as many of your own questions as possible before proceeding. Example 1-3. Retrieving Twitter search trends >>> import twitter >>> twitter_search = twitter.Twitter(domain="search.twitter.com") >>> trends = twitter_search.trends() >>> [ trend['name'] for trend in trends['trends'] ] [u'#ZodiacFacts', u'#nowplaying', u'#ItsOverWhen', u'#Christoferdrew', u'Justin Bieber', u'#WhatwouldItBeLike', u'#Sagittarius', u'SNL', u'#SurveySays', u'#iDoit2'] Since you’re probably wondering, the pattern for using the twitter module is simple and predictable: instantiate the Twitter class with a base URL and then invoke methods on the object that correspond to URL contexts. For example, twit ter_search.trends() initiates an HTTP call to GET http://search.twitter.com/trends .json, which you could type into your web browser to get the same set of results. As further context for the previous interpreter session, this chapter was originally drafted on a Saturday night, so it’s not a coincidence that the trend SNL (Saturday Night Live, a popular comedy show that airs in the United States) appears in the list. Now might be a good time to go ahead and bookmark the official Twitter API documenta- tion since you’ll be referring to it quite frequently. Given that SNL is trending, the next logical step might be to grab some search results about it by using the search API to search for tweets containing that text and then print them out in a readable way as a JSON structure. Example 1-4 illustrates. Example 1-4. Paging through Twitter search results >>> search_results = [] >>> for page in range(1,6): ... search_results.append(twitter_search.search(q="SNL", rpp=100, page=page)) The code fetches and stores five consecutive batches (pages) of results for a query (q) of SNL, with 100 results per page (rpp). It’s again instructive to observe that the equiv- alent REST query‡ that we execute in the loop is of the form http://search.twitter.com/ search.json?&q=SNL&rpp=100&page=1. The trivial mapping between the REST API and the twitter module makes it very simple to write Python code that interacts with Twitter services. After executing the search, the search_results list contains five objects, each of which is a batch of 100 results. You can print out the results in a readable way for inspection by using the json package that comes built-in as of Python Version 2.6, as shown in Example 1-5. ‡ If you’re not familiar with REST, see the sidebar “RESTful Web Services” on page 49 in Chapter 7 for a brief explanation. Collecting and Manipulating Twitter Data | 5 Example 1-5. Pretty-printing Twitter data as JSON >>> import json >>> print json.dumps(search_results, sort_keys=True, indent=1) [ { "completed_in": 0.088122000000000006, "max_id": 11966285265, "next_page": "?page=2&max_id=11966285265&rpp=100&q=SNL", "page": 1, "query": "SNL", "refresh_url": "?since_id=11966285265&q=SNL", "results": [ { "created_at": "Sun, 11 Apr 2010 01:34:52 +0000", "from_user": "bieber_luv2", "from_user_id": 106998169, "geo": null, "id": 11966285265, "iso_language_code": "en", "metadata": { "result_type": "recent" }, "profile_image_url": "http://a1.twimg.com/profile_images/809471978/DSC00522...", "source": "<a href="http://twitter.com/">web</a>", "text": " ...truncated... im nt gonna go to sleep happy unless i see @justin...", "to_user_id": null } ... output truncated - 99 more tweets ... ], "results_per_page": 100, "since_id": 0 }, ... output truncated - 4 more pages ... ] Be advised that as of late 2010, the from_user_id field in each search result does not correspond to the tweet author’s actual Twitter id. See Twitter API Issue #214 for details. This defect is of no consequence to example code in this chapter, but it is important to note in case you start getting creative (which is highly encouraged). We’ll wait until later in the book to pick apart many of the details in this query (see Chapter 5); the important observation at the moment is that the tweets are keyed by results in the response. We can distill the text of the 500 tweets into a list with the following approach. Example 1-6 illustrates a double list comprehension that’s inden- ted to illustrate the intuition behind it being nothing more than a nested loop. 6 | Chapter 1: Introduction: Hacking on Twitter Data Example 1-6. A simple list comprehension in Python >>> tweets = [ r['text'] \ ... for result in search_results \ ... for r in result['results'] ] List comprehensions are used frequently throughout this book. Although they can look quite confusing if written on a single line, printing them out as nested loops clarifies the meaning. The result of tweets in this particular case is equivalent to defining an empty list called tweets and invoking tweets.append(r['text']) in the same kind of nested loop as presented here. See the “Data Structures” section in the official Python tutorial for more details. List comprehensions are particularly powerful because they usually yield substantial performance gains over nested lists and provide an intuitive (once you’re familiar with them) yet terse syntax. Frequency Analysis and Lexical Diversity One of the most intuitive measurements that can be applied to unstructured text is a metric called lexical diversity. Put simply, this is an expression of the number of unique tokens in the text divided by the total number of tokens in the text, which are elementary yet important metrics in and of themselves. It could be computed as shown in Exam- ple 1-7. Example 1-7. Calculating lexical diversity for tweets >>> words = [] >>> for t in tweets: ... words += [ w for w in t.split() ] ... >>> len(words) # total words 7238 >>> len(set(words)) # unique words 1636 >>> 1.0*len(set(words))/len(words) # lexical diversity 0.22602928985907708 >>> 1.0*sum([ len(t.split()) for t in tweets ])/len(tweets) # avg words per tweet 14.476000000000001 Prior to Python 3.0, the division operator applies the floor function and returns an integer value (unless one of the operands is a floating-point value). Multiply either the numerator or the denominator by 1.0 to avoid truncation errors. One way to interpret a lexical diversity of around 0.23 would be to say that about one out of every four words in the aggregated tweets is unique. Given that the average number of words in each tweet is around 14, that translates to just over 3 unique words per tweet. Without introducing any additional information, that could be interpreted as meaning that each tweet carries about 20 percent unique information. What would be interesting to know at this point is how “noisy” the tweets are with uncommon Collecting and Manipulating Twitter Data | 7 abbreviations users may have employed to stay within the 140 characters, as well as what the most frequent and infrequent terms used in the tweets are. A distribution of the words and their frequencies would be helpful. Although these are not difficult to compute, we’d be better off installing a tool that offers a built-in frequency distribution and many other tools for text analysis. The Natural Language Toolkit (NLTK) is a popular module we’ll use throughout this book: it delivers a vast amount of tools for various kinds of text analytics, including the calculation of common metrics, information extraction, and natural language process- ing (NLP). Although NLTK isn’t necessarily state-of-the-art as compared to ongoing efforts in the commercial space and academia, it nonetheless provides a solid and broad foundation—especially if this is your first experience trying to process natural language. If your project is sufficiently sophisticated that the quality or efficiency that NLTK provides isn’t adequate for your needs, you have approximately three options, de- pending on the amount of time and money you are willing to put in: scour the open source space for a suitable alternative by running comparative experiments and bench- marks, churn through whitepapers and prototype your own toolkit, or license a com- mercial product. None of these options is cheap (assuming you believe that time is money) or easy. NLTK can be installed per the norm with easy_install, but you’ll need to restart the interpreter to take advantage of it. You can use the cPickle module to save (“pickle”) your data before exiting your current working session, as shown in Example 1-8. Example 1-8. Pickling your data >>> f = open("myData.pickle", "wb") >>> import cPickle >>> cPickle.dump(words, f) >>> f.close() >>> $ easy_install nltk Searching for nltk ...truncated output... Finished processing dependencies for nltk If you encounter an “ImportError: No module named yaml” problem when you try to import nltk, execute an easy_install pyYaml, which should clear it up. After installing NLTK, you might want to take a moment to visit its official website, where you can review its documentation. This includes the full text of Steven Bird, Ewan Klein, and Edward Loper’s Natural Language Processing with Python (O’Reilly), NLTK’s authoritative reference. 8 | Chapter 1: Introduction: Hacking on Twitter Data What are people talking about right now? Among the most compelling reasons for mining Twitter data is to try to answer the question of what people are talking about right now. One of the simplest techniques you could apply to answer this question is basic frequency analysis. NLTK simplifies this task by providing an API for frequency analysis, so let’s save ourselves some work and let NLTK take care of those details. Example 1-9 demonstrates the findings from creating a frequency distribution and takes a look at the 50 most frequent and least frequent terms. Example 1-9. Using NLTK to perform basic frequency analysis >>> import nltk >>> import cPickle >>> words = cPickle.load(open("myData.pickle")) >>> freq_dist = nltk.FreqDist(words) >>> freq_dist.keys()[:50] # 50 most frequent tokens [u'snl', u'on', u'rt', u'is', u'to', u'i', u'watch', u'justin', u'@justinbieber', u'be', u'the', u'tonight', u'gonna', u'at', u'in', u'bieber', u'and', u'you', u'watching', u'tina', u'for', u'a', u'wait', u'fey', u'of', u'@justinbieber:', u'if', u'with', u'so', u"can't", u'who', u'great', u'it', u'going', u'im', u':)', u'snl...', u'2nite...', u'are', u'cant', u'dress', u'rehearsal', u'see', u'that', u'what', u'but', u'tonight!', u':d', u'2', u'will'] >>> freq_dist.keys()[-50:] # 50 least frequent tokens [u'what?!', u'whens', u'where', u'while', u'white', u'whoever', u'whoooo!!!!', u'whose', u'wiating', u'wii', u'wiig', u'win...', u'wink.', u'wknd.', u'wohh', u'won', u'wonder', u'wondering', u'wootwoot!', u'worked', u'worth', u'xo.', u'xx', u'ya', u'ya<3miranda', u'yay', u'yay!', u'ya\u2665', u'yea', u'yea.', u'yeaa', u'yeah!', u'yeah.', u'yeahhh.', u'yes,', u'yes;)', u'yess', u'yess,', u'you!!!!!', u"you'll", u'you+snl=', u'you,', u'youll', u'youtube??', u'youu<3', u'youuuuu', u'yum', u'yumyum', u'~', u'\xac\xac'] Python 2.7 added a collections.Counter (http://docs.python.org/library/ collections.html#collections.Counter) class that facilitates counting op- erations. You might find it useful if you’re in a situation where you can’t easily install NLTK, or if you just want to experiment with the latest and greatest classes from Python’s standard library. A very quick skim of the results from Example 1-9 shows that a lot more useful infor- mation is carried in the frequent tokens than the infrequent tokens. Although some work would need to be done to get a machine to recognize as much, the frequent tokens refer to entities such as people, times, and activities, while the infrequent terms amount to mostly noise from which no meaningful conclusion could be drawn. The first thing you might have noticed about the most frequent tokens is that “snl” is at the top of the list. Given that it is the basis of the original search query, this isn’t surprising at all. Where it gets more interesting is when you skim the remaining tokens: there is apparently a lot of chatter about a fellow named Justin Bieber, as evidenced by Collecting and Manipulating Twitter Data | 9 the tokens @justinbieber, justin, and bieber. Anyone familiar with SNL would also know that the occurrences of the tokens “tina” and “fey” are no coincidence, given Tina Fey’s longstanding affiliation with the show. Hopefully, it’s not too difficult (as a human) to skim the tokens and form the conjecture that Justin Bieber is a popular guy, and that a lot of folks were very excited that he was going to be on the show on the Saturday evening the search query was executed. At this point, you might be thinking, “So what? I could skim a few tweets and deduce as much.” While that may be true, would you want to do it 24/7, or pay someone to do it for you around the clock? And what if you were working in a different domain that wasn’t as amenable to skimming random samples of short message blurbs? The point is that frequency analysis is a very simple, yet very powerful tool that shouldn’t be overlooked just because it’s so obvious. On the contrary, it should be tried out first for precisely the reason that it’s so obvious and simple. Thus, one preliminary takeaway here is that the application of a very simple technique can get you quite a long way toward answering the question, “What are people talking about right now?” As a final observation, the presence of “rt” is also a very important clue as to the nature of the conversations going on. The token RT is a special symbol that is often prepended to a message to indicate that you are retweeting it on behalf of someone else. Given the high frequency of this token, it’s reasonable to infer that there were a large amount of duplicate or near-duplicate tweets involving the subject matter at hand. In fact, this observation is the basis of our next analysis. The token RT can be prepended to a message to indicate that it is being relayed, or “retweeted” in Twitter parlance. For example, a tweet of “RT @SocialWebMining Justin Bieber is on SNL 2nite. w00t?!?” would in- dicate that the sender is retweeting information gained via the user @SocialWebMining. An equivalent form of the retweet would be “Justin Bieber is on SNL 2nite. w00t?!? Ummm…(via @SocialWebMining)”. Extracting relationships from the tweets Because the social web is first and foremost about the linkages between people in the real world, one highly convenient format for storing social web data is a graph. Let’s use NetworkX to build out a graph connecting Twitterers who have retweeted infor- mation. We’ll include directionality in the graph to indicate the direction that infor- mation is flowing, so it’s more precisely called a digraph. Although the Twitter APIs do offer some capabilities for determining and analyzing statuses that have been retweeted, these APIs are not a great fit for our current use case because we’d have to make a lot of API calls back and forth to the server, which would be a waste of the API calls included in our quota. 10 | Chapter 1: Introduction: Hacking on Twitter Data At the time this book was written, Twitter imposes a rate limit of 350 API calls per hour for authenticated requests; anonymous requests are limited to 150 per hour. You can read more about the specifics at http: //dev.twitter.com/pages/rate-limiting. In Chapters 4 and 5, we’ll discuss techniques for making the most of the rate limiting, as well as some other creative options for collecting data. Besides, we can use the clues in the tweets themselves to reliably extract retweet infor- mation with a simple regular expression. By convention, Twitter usernames being with an @ symbol and can only include letters, numbers, and underscores. Thus, given the conventions for retweeting, we only have to search for the following patterns: • RT followed by a username • via followed by a username Although Chapter 5 introduces a module specifically designed to parse entities out of tweets, Example 1-10 demonstrates that you can use the re module to compile§ a pat- tern and extract the originator of a tweet in a lightweight fashion, without any special libraries. Example 1-10. Using regular expressions to find retweets >>> import re >>> rt_patterns = re.compile(r"(RT|via)((?:\b\W*@\w+)+)", re.IGNORECASE) >>> example_tweets = ["RT @SocialWebMining Justin Bieber is on SNL 2nite. w00t?!?", ... "Justin Bieber is on SNL 2nite. w00t?!? (via @SocialWebMining)"] >>> for t in example_tweets: ... rt_patterns.findall(t) ... [('RT', ' @SocialWebMining')] [('via', ' @SocialWebMining')] In case it’s not obvious, the call to findall returns a list of tuples in which each tuple contains either the matching text or an empty string for each group in the pattern; note that the regex does leave a leading space on the extracted entities, but that’s easily fixed with a call to strip(), as demonstrated in Example 1-11. Since neither of the example tweets contains both of the groups enclosed in the parenthetical expressions, one string is empty in each of the tuples. Regular expressions are a basic programming concept whose explana- tion is outside the scope of this book. The re module documentation is a good place to start getting up to speed, and you can always consult Friedl’s classic Mastering Regular Expressions (O’Reilly) if you want to learn more than you’ll probably ever need to know about them. § In the present context, compiling a regular expression means transforming it into bytecode so that it can be executed by a matching engine written in C. Collecting and Manipulating Twitter Data | 11 D o wnload from Wow! eBook Given that the tweet data structure as returned by the API provides the username of the person tweeting and the newly found ability to extract the originator of a retweet, it’s a simple matter to load this information into a NetworkX graph. Let’s create a graph in which nodes represent usernames and a directed edge between two nodes signifies that there is a retweet relationship between the nodes. The edge itself will carry a pay- load of the tweet ID and tweet text itself. Example 1-11 demonstrates the process of generating such a graph. The basic steps involved are generalizing a routine for extracting usernames in retweets, flattening out the pages of tweets into a flat list for easier processing in a loop, and finally, iterating over the tweets and adding edges to a graph. Although we’ll generate an image of the graph later, it’s worthwhile to note that you can gain a lot of insight by analyzing the characteristics of graphs without necessarily visualizing them. Example 1-11. Building and analyzing a graph describing who retweeted whom >>> import networkx as nx >>> import re >>> g = nx.DiGraph() >>> >>> all_tweets = [ tweet ... for page in search_results ... for tweet in page["results"] ] >>> >>> def get_rt_sources(tweet): ... rt_patterns = re.compile(r"(RT|via)((?:\b\W*@\w+)+)", re.IGNORECASE) ... return [ source.strip() ... for tuple in rt_patterns.findall(tweet) ... for source in tuple ... if source not in ("RT", "via") ] ... >>> for tweet in all_tweets: ... rt_sources = get_rt_sources(tweet["text"]) ... if not rt_sources: continue ... for rt_source in rt_sources: ... g.add_edge(rt_source, tweet["from_user"], {"tweet_id" : tweet["id"]}) ... >>> g.number_of_nodes() 160 >>> g.number_of_edges() 125 >>> g.edges(data=True)[0] (u'@ericastolte', u'bonitasworld', {'tweet_id': 11965974697L}) >>> len(nx.connected_components(g.to_undirected())) 37 >>> sorted(nx.degree(g)) [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 4, 4, 4, 5, 6, 6, 9, 37] 12 | Chapter 1: Introduction: Hacking on Twitter Data The built-in operations that NetworkX provides are a useful starting point to make sense of the data, but it’s important to keep in mind that we’re only looking at a very small slice of the overall conversation happening on Twitter about SNL—500 tweets out of potentially tens of thousands (or more). For example, the number of nodes in the graph tells us that out of 500 tweets, there were 160 users involved in retweet relationships with one another, with 125 edges connecting those nodes. The ratio of 160/125 (approximately 1.28) is an important clue that tells us that the average de- gree of a node is approximately one—meaning that although some nodes are connected to more than one other node, the average is approximately one connection per node. The call to connected_components shows us that the graph consists of 37 subgraphs and is not fully connected. The output of degree might seem a bit cryptic at first, but it actually confirms insight we’ve already gleaned: think of it as a way to get the gist of how well connected the nodes in the graph are without having to render an actual graph. In this case, most of the values are 1, meaning all of those nodes have a degree of 1 and are connected to only one other node in the graph. A few values are between 2 and 9, indicating that those nodes are connected to anywhere between 2 and 9 other nodes. The extreme outlier is the node with a degree of 37. The gist of the graph is that it’s mostly composed of disjoint nodes, but there is one very highly connected node. Figure 1-1 illustrates a distribution of degree as a column chart. The trendline shows that the distribution closely follows a Power Law and has a “heavy” or “long” tail. Although the characteristics of distributions with long tails are by no means treated with rigor in this book, you’ll find that lots of distributions we’ll encounter exhibit this property, and you’re highly encouraged to take the initiative to dig deeper if you feel the urge. A good starting point is Zipf’s law. Figure 1-1. A distribution illustrating the degree of each node in the graph, which reveals insight into the graph’s connectedness Collecting and Manipulating Twitter Data | 13 We’ll spend a lot more time in this book using automatable heuristics to make sense of the data; this chapter is intended simply as an introduction to rattle your brain and get you thinking about ways that you could exploit data with the low-hanging fruit that’s available to you. Before we wrap up this chapter, however, let’s visualize the graph just to be sure that our intuition is leading us in the right direction. Visualizing Tweet Graphs Graphviz is a staple in the visualization community. This section introduces one pos- sible approach for visualizing graphs of tweet data: exporting them to the DOT lan- guage, a simple text-based format that Graphviz consumes. Graphviz binaries for all platforms can be downloaded from its official website, and the installation is straight- forward regardless of platform. Once Graphviz is installed, *nix users should be able to easy_install pygraphviz per the norm to satisfy the PyGraphviz dependency Net- workX requires to emit DOT. Windows users will most likely experience difficulties installing PyGraphviz,‖ but this turns out to be of little consequence since it’s trivial to tailor a few lines of code to generate the DOT language output that we need in this section. Example 1-12 illustrates an approach that works for both platforms. Example 1-12. Generating DOT language output is easy regardless of platform OUT = "snl_search_results.dot" try: nx.drawing.write_dot(g, OUT) except ImportError, e: # Help for Windows users: # Not a general-purpose method, but representative of # the same output write_dot would provide for this graph # if installed and easy to implement dot = ['"%s" -> "%s" [tweet_id=%s]' % (n1, n2, g[n1][n2]['tweet_id']) \ for n1, n2 in g.edges()] f = open(OUT, 'w') f.write('strict digraph {\n%s\n}' % (';\n'.join(dot),)) f.close() The DOT output that is generated is of the form shown in Example 1-13. ‖ See NetworkX Ticket #117, which reveals that this has been a long-standing issue that somehow has not garnered the support to be overcome even after many years of frustration. The underlying issue has to do with the need to compile C code during the easy_install process. The ability to work around this issue fairly easily by generating DOT language output may be partly responsible for why it has remained unresolved for so long. 14 | Chapter 1: Introduction: Hacking on Twitter Data Example 1-13. Example DOT language output strict digraph { "@ericastolte" -> "bonitasworld" [tweet_id=11965974697]; "@mpcoelho" -> "Lil_Amaral" [tweet_id=11965954427]; "@BieberBelle123" -> "BELIEBE4EVER" [tweet_id=11966261062]; "@BieberBelle123" -> "sabrina9451" [tweet_id=11966197327]; } With DOT language output on hand, the next step is to convert it into an image. Graphviz itself provides a variety of layout algorithms to visualize the exported graph; circo, a tool used to render graphs in a circular-style layout, should work well given that the data suggested that the graph would exhibit the shape of an ego graph with a “hub and spoke”-style topology, with one central node being highly connected to many nodes having a degree of 1. On a *nix platform, the following command converts the snl_search_results.dot file exported from NetworkX into an snl_search_results.dot.png file that you can open in an image viewer (the result of the operation is displayed in Figure 1-2): $ circo -Tpng -Osnl_search_results snl_search_results.dot Windows users can use the GVedit application to render the file as shown in Fig- ure 1-3. You can read more about the various Graphviz options in the online docu- mentation. Visual inspection of the entire graphic file confirms that the characteristics of the graph align with our previous analysis, and we can visually confirm that the node with the highest degree is @justinbieber, the subject of so much discussion (and, in case you missed that episode of SNL, the guest host of the evening). Keep in mind that if we had harvested a lot more tweets, it is very likely that we would have seen many more interconnected subgraphs than are evidenced in the sampling of 500 tweets that we have been analyzing. Further analysis of the graph is left as a voluntary exercise for the reader, as the primary objective of this chapter was to get your development envi- ronment squared away and whet your appetite for more interesting topics. Graphviz appears elsewhere in this book, and if you consider yourself to be a data scientist (or are aspiring to be one), it is a tool that you’ll want to master. That said, we’ll also look at many other useful approaches to visualizing graphs. In the chapters to come, we’ll cover additional outlets of social web data and techniques for analysis. Synthesis: Visualizing Retweets with Protovis A turn-key example script that synthesizes much of the content from this chapter and adds a visualization is how we’ll wrap up this chapter. In addition to spitting some useful information out to the console, it accepts a search term as a command line pa- rameter, fetches, parses, and pops up your web browser to visualize the data as an interactive HTML5-based graph. It is available through the official code repository for this book at http://github.com/ptwobrussell/Mining-the-Social-Web/blob/master/python _code/introduction__retweet_visualization.py. You are highly encouraged to try it out. Collecting and Manipulating Twitter Data | 15 We’ll revisit Protovis, the underlying visualization toolkit for this example, in several chapters later in the book. Figure 1-4 illustrates Protovis output from this script. The boilerplate in the sample script is just the beginning—much more can be done! Figure 1-2. Our search results rendered in a circular layout with Graphviz 16 | Chapter 1: Introduction: Hacking on Twitter Data D o wnload from Wow! eBook Figure 1-3. Windows users can use GVedit instead of interacting with Graphviz at the command prompt Closing Remarks This chapter got you up and running, and illustrated how easy it is to use Python’s interactive interpreter to explore and visualize Twitter data. Before you move on to other chapters, it’s important that you feel comfortable with your Python development environment, and it’s highly recommended that you spend some time with the Twitter APIs and Graphviz. If you feel like going out on a tangent, you might want to check out canviz, a project that aims to draw Graphviz graphs on a web browser