Scraping Ice and Fire

Posted on May 26, 2016

Contributed by Rob Castellano. He is currently in the NYC Data Science Academy 12 week full time Data Science Bootcamp program taking  place between April 11th to July 1st, 2016. This post is based on his third class project - Python web scraping (due on the 6th week of the program).

A Song of Ice and Fire (ASOIAF) is a series of fantasy novels by George R. R. Martin. It is also the basis of the hit HBO television show Game of Thrones. The world of ASOIAF is large, complex, and has a lot of source material (and hopefully more to come!). As a big fan of the books and the television show, I was interested in examining data and exploring various facets of the series. I was also curious how my opinions on aspects of the series (such as favorite books and characters) compared to other fans.

This blog post is written to be readable for anyone, however those who have read the books or watch the television show will probably gain more insight than those who have not. Also, SPOILER WARNING! Information in this post can potentially spoil material in the books and television've been warned.

The code for this project can be found on GitHub. An overview of the web scraping required for this project is included at the end of this post.


The popularity of the ASOIAF series and Game of Thrones has resulted in a large community of fans. For this project, I scraped data from, a self-described "Encyclopedia of Ice and Fire." It is a fan-created reference for both the books and the television show with information such as:

  • Character lists and information
  • Family trees
  • Maps
  • Historical timeline
  • Chapter guides
  • Episode guides

Before describing the data I gathered, I will give some basic information about the series.

About the ASOIAF series

There are five books in the series (so far):

  1. A Game of Thrones (AGOT)
  2. A Clash of Kings (ACOK)
  3. A Storm of Swords (ASOS)
  4. A Feast for Crows (AFFC)
  5. A Dance with Dragons (ADWD)

I will use these abbreviations for each book. The average length of each book is almost 1000 pages. There are three main storylines that intertwine throughout the novels. There are many different settings and an extremely large collection of characters, each belonging to a "house" (family), and many complex relationships among them.

One of the narrative features of ASOIAF that I found most interesting and lends itself to analysis is that each chapter has a point of view (POV) character. Within a chapter, the action is narrated in the limited third person from the perspective of the POV character. Thus, the reader develops relationships with several characters; there is not one main character and characters you think are main characters die. This narrative structure results in different chapters having different styles and readers having a wide range of favorite characters.

Data to be analyzed contains  chapter summaries for every chapter. For each chapter I gathered the following information:

  • Book
  • Chapter number within the book
  • POV character
  • Chapter name (sometimes just the POV character name, sometimes different)
  • Chapter blurb (one sentence description of the chapter)
  • Summary (a few paragraphs describing the contents of the chapter)
  • Characters appearing in the chapter
  • Score of the chapter (voted on by users on a scale of 1 to 10)

The data is gathered by members of the TowerOfTheHand community. Note also that not all chapters of A Dance with Dragons (the most recent book) have summaries--summary writing is a work in progress. At the end of this post, in the Scraping section, I describe how I scraped this data.

Initial questions

The following were the questions I was interested in exploring in the data:

  1. What is the best book?
  2. What are the community's favorite characters?
  3. How do my opinions on books and characters compare to the TowerOfTheHand community?
  4. Do books and characters have arcs?
  5. Organize the data in such a way that users can explore data on their favorite characters and chapters.

Chapter score data

All scores

There are 344 chapters is the series and users on TowerOfTheHand rate each chapter on a scale of 1 to 10. The mean score is 7.9, with the minimum of 5.84 and a maximum of 9.32. Below we have the distribution of scores across all chapters.


Scores by book

I next looked at the distribution of scores by book. The books ordered from highest to lowest mean chapter score is: AGOT, ASOS, ADWD, ACOK, AFFC. AGOT also has the second lowest standard deviation, meaning it is a highly rated and consistent book. My personal rankings of the books would have AGOT and ASOS switched; I was surprised that AGOT was voted higher and more consistent than ASOS as ASOS is my favorite book and the favorite of many people I've talked to. ADWD being good but inconsistent agrees with my opinion. More information on books will be seen in a Shiny app below.


Mean Standard Deviation
A Game of Thrones 8.22 .57
A Clash of Kings 7.75 .71
A Storm of Swords 8.03 .65
A Feast for Crows 7.57 .52
A Dance with Dragons 8.00 .71

POV Characters

There are 31 POV characters in the series. Nine of these POV characters are one-off POV characters and several more have only a few POV chapters.


We can focus on what I call the "main" POV characters; these are characters with several POV chapters and all are characters that viewers of Game of Thrones would be familiar with. The character with the most POV chapters is Tyrion Lannister. Four of the seven characters with the most POV chapters are Stark children, with two more of the top nine being the Stark parents. Women are six of the fourteen main POV characters.


I next looked at chapter scores for each POV character. We can see that several of the top-rated characters are from one-off POV chapters. However, the three lowest rated characters are also infrequent POV characters.

Again focusing on the main POV characters, Ned Stark and Theon Greyjoy have the best, most consistently highly rated chapters. The six lowest rated characters are the women main POV characters. Daenerys Targaryen, despite being one of the main characters of the series, has an extremely large range of scores for her chapters.


App for POV characters and chapter scores

I created a Shiny app to help visualize all chapter scores. Explore this app here.

In this app you can display all of the chapter scores, view the ratings trends for each book, and highlight the chapters of any number of POV characters. Note that hovering over a chapter will show the POV character, chapter name, book, and score for that chapter.

By looking at all the chapter scores and the trends for each books we can see: AGOT is consistently highly rated, ACOK is not as highly rated and dips towards the end of the book,  ASOS climbs in ratings and ends in a huge climax, AFFC gains momentum as the book progresses, and ADWD is highly rated, but with large variability. This agrees with some of my opinions of each book. For example, although AGOT may be rated higher overall, the end of ASOS is incredibly exciting. Also AFFC starts slow, but by the end is very good and on par with the other books.


We can also select POV characters to display. In the simplest example below, we selected Daenerys Targaryen and her chapters appear. We can see that she has several good chapters in AGOT, ending in a great chapter (dragons!), her appearances in ACOK are limited, but she has one great chapter, she also has consistently good chapters in ASOS, but her chapters in ADWD are among the worst, save one chapter.


I hope you will use this app to explore your favorite characters, discover insights on character arcs, and relive your favorite chapters!

Word clouds

I created word clouds for many of the main characters/houses in ASOIAF based on the summaries from POV chapters from that character/house and in the style of their house banner.

All chapter summaries



Jon Snow


Tyrion Lannister



Daenerys Targaryen



Catelyn Tully



Theon, Asha, and Aeron Greyjoy


Arianne Martell



Arya Stark


Brandon Stark


Further Directions

This analysis is just the beginning of what could be done with the ASOIAF data I scraped. Some other areas I would like to explore are:

  1. Analyze by book quantiles (thirds, quarter, etc.). Do certain portions of the books tend to be better than others?
  2. There is much more natural language processing that can be done with the chapter summaries (beyond the word clouds I produced).
  3. Explore the relationships between characters using the information on characters appearing in each chapter. There are 1111 total characters appearing! How do they relate to one another? Which characters are most connected and least connected?


As mentioned earlier, I obtained the data I used from by scraping information on each chapter. All of the code to do this can be found in the "scraping" folder on  GitHub.

To scrape, I used the Python package Scrapy. Briefly, Scrapy is a package that facilitates extraction of information from well structured websites. You specify the type of information you want, what to do with that information, and how to get it. The portion that describes how to get the data is called the "spider" because it crawl though the website (and sometimes from one website to another) collecting the data you've specified.

Below I've included the code for the spider that scraped the data for this project. This is most of the hard work that is involved in getting the data. I then exported this data to MongoDB from which one can analyze, write to a .csv file, etc. Worth noting is that I also use the Selenium package, which simulates a browser to help facilitate scraping the data. This is necessary because the score of each chapter is dynamically fetched based on current voting.

About Author

Rob Castellano

Rob recently received his Ph.D. in Mathematics from Columbia. His training as a pure mathematician has given him strong quantitative skills and experience in using creative problem solving techniques. He has experience conveying abstract concepts to both experts...
View all posts by Rob Castellano >

Related Articles

Leave a Comment

No comments found.

View Posts by Categories

Our Recent Popular Posts

View Posts by Tags

#python #trainwithnycdsa 2019 2020 Revenue 3-points agriculture air quality airbnb airline alcohol Alex Baransky algorithm alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus ames dataset ames housing dataset apartment rent API Application artist aws bank loans beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep boston safety Bundles cake recipe California Cancer Research capstone car price Career Career Day citibike classic cars classpass clustering Coding Course Demo Course Report covid 19 credit credit card crime frequency crops D3.js data data analysis Data Analyst data analytics data for tripadvisor reviews data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization database Deep Learning Demo Day Discount disney dplyr drug data e-commerce economy employee employee burnout employer networking environment feature engineering Finance Financial Data Science fitness studio Flask flight delay gbm Get Hired ggplot2 googleVis H20 Hadoop hallmark holiday movie happiness healthcare frauds higgs boson Hiring hiring partner events Hiring Partners hotels housing housing data housing predictions housing price hy-vee Income Industry Experts Injuries Instructor Blog Instructor Interview insurance italki Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter las vegas airport lasso regression Lead Data Scienctist Lead Data Scientist leaflet league linear regression Logistic Regression machine learning Maps market matplotlib Medical Research Meet the team meetup methal health miami beach movie music Napoli NBA netflix Networking neural network Neural networks New Courses NHL nlp NYC NYC Data Science nyc data science academy NYC Open Data nyc property NYCDSA NYCDSA Alumni Online Online Bootcamp Online Training Open Data painter pandas Part-time performance phoenix pollutants Portfolio Development precision measurement prediction Prework Programming public safety PwC python Python Data Analysis python machine learning python scrapy python web scraping python webscraping Python Workshop R R Data Analysis R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn seafood type Selenium sentiment analysis sentiment classification Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau teachers team team performance TensorFlow Testimonial tf-idf Top Data Science Bootcamp Top manufacturing companies Transfers tweets twitter videos visualization wallstreet wallstreetbets web scraping Weekend Course What to expect whiskey whiskeyadvocate wildfire word cloud word2vec XGBoost yelp youtube trending ZORI