What do people want for Christmas? To figure this out, we've hooked a computer program up to the Twitter stream. It grabs any tweets containing a Christmas wish. Here's a breakdown of all those wishes. These results are updated every 20 minutes, so check back to see how Christmas wishing changes as the big day approaches!
|Days to Christmas||0
The Xmas Tweet project is a product of the RuthsResearch Lab in the School of Computer Science at McGill University. This project is part of a larger initiative to develop techniques to mine individual attributes and activities from unstructured, online data. The system and website were built by Derek Ruths. Mathieu Perreault contributed the scrolling tweets at the top of the page.
We'd love to hear your feedback and ideas!
How It Works
Tweets were obtained from the public Twitter stream. This stream provided us with approximately 1% of all tweets containing the word "Christmas". Clearly, not all tweets containing the word "Christmas" contain wishes. As a result, a step (discussed next) was required to identify and extract wishes contained in tweets.
Extracting Christmas Wishes
We extracted wishes from all tweets containing the phrases:
- "I want [a/the] ________ for Christmas"
- "For Christmas I want [a/the] ________"
Retweets were discarded. The valid wishes were then further cleaned to remove common misspellings and replace abbreviations with complete words.
The ranking shown in the "Top Wishes" tab was constructed by counting the number of times a given wish was made.
Getting the computer to perform accurate categorization of words is very hard. In this project, the challenge was compounded by the unrestricted vocabulary individuals could use to specify their wishes. In order to select meaningful categories and categorize as many of these words as possible, we had our computer algorithm learn the categorization scheme of Craig's List. The items being sold in the different categories provided enough information to classify most wishes being made.
This is a list of the most-tweeted Christmas wishes. The bar chart on the right shows the number of tweet wishes for that specific item. Note that only wishes that used very similar wording were grouped together (e.g. wishes for "money" and "cash" would fall into different categories). Because this can be a bit misleading, we added the "By Category" tab which shows all items grouped by category.
The most wished for item was "you". It received so many tweets, in fact, that displaying it with other top wishes would completely squash the bar plot below. As a result, the top wish results including "you" are shown on this separate page.
- Justin Bieber is a hot commodity this Christmas.
- A number of song lyrics appear in the wish list.
- Apparently more women are wishing for romance than men: "him" and "boyfriend" rank much higher than "her".
- Dogs/puppies are a wish list favorite.
The Wish List
Here we provide a breakdown of the top wishes in various categories.
This list shows the number of wishes that mentioned brand names (not their products).
Apple vs. Android
In this plot, all wishes have been grouped into categories. The size of the bar indicates the number of wishes received for items in that category. Because of their overwhelming popularity, Justin Bieber and hippopotamuses have been given their own categories.