Archive for September, 2010
Yahoo! Releases Timeline Prototype: How Does it Stack Up?
Some exciting news on the Twitter wire last week as Yahoo released a prototype of an interesting news search engine that “generates timelines that stretch into the future as well as the past.” Developed by Yahoo’s Barcelona Research Lab, the new search service, Time Explorer, visually displays the frequency and volume of news events on a Simile timeline.
Perhaps the most striking feature of Time Explorer is the ability to search in the past AND the future. Yahoo’s goal is a lofty one: essentially to build a news search engine that predicts future outcomes of today’s issues using predictions found in the text of past articles. For example, I searched for World Cup in the past and the site came back with 456,000 plus results with the word “World Cup.” Search the same term in the future and you are directed to articles that make predictions of where the upcoming World Cups will be held, or who the US Soccer coach will be in 2014, very cool stuff.
Since Yahoo has decided to dip their toe into the business of interactive data visualization I wanted to take a closer look at how Dipity and Time Explorer stack up. Since Time Explorer is relatively new and still under development, I decided to even the playing field a bit. It would be fruitless and perhaps unfair to compare the two based on the scope of applications, user-friendliness, visual aesthetic or integration capabilities. As the original and most widely used visual timeline site, Dipity has a clear and expected advantage on those fronts. However, let’s take a closer look on some fronts where Time Explorer can compete, specifically search capabilities, news aggregation an the “serendipity factor.”
Search Capabilities
Time Explorer is built from the data of 1.8 million New York Times articles published over a twenty year period. Within this limited data set, Time Explorer’s search capabilities seem to be extremely effective. Search results are sorted based on the popularity of each article and offer keywords that you can adjust to further refine your search. For example, if searching for “Global Warming” produces an article with the keyword “Arnold Schwarzenegger,” users can click the name to search this keyword, see more or less articles with the keyword, or even exclude the term from their search. Time Explorer also offers users information on each article’s popularity, source and author. In addition, users can select from searching the past or searching for the future. The future search function looks for articles that discuss future predictions for specific topics.

In the visual portion of the search, users can scan specific dates for articles by scrolling through the area at the bottom of the screen. This helps to quickly shift through time periods in order to look for articles from a specific time period. I find this to me the most useful feature of the visual aspect of Time Explorer.

Dipity, unfortunately, does not have a search engine powered by Yahoo!. However, Dipity users can search through a much larger amount and type of information. Instead of offering users access to a preexisting database, Dipity allows users to use existing search APIs from other sites (Google News, Newsline, YouTube, Twitter, Flickr, etc.) in order to gather information and media and create their own topic. For example, in order to create a timeline about Tiger Woods, I might try integrating a Google News search using the terms “Tiger Woods Golf Career,” and then add “Tiger Woods Affair.” Using these two separate search terms will yield a more complete topic based around the subject, Tiger Woods. We can also use a YouTube search to add in some videos of that sweet swing, or a news report or two about the upcoming divorce.

In addition, Dipity itself offers over 800,000 preexisting user-created topics to search through. Each one of these topics feature multiple events, full of text, photos, videos, music and links and more. In fact, there are so many events on Dipity, that it is tough to even offer an estimate of just how extensive the database of information is.

My favorite feature of Time Explorer is the ability to visually scroll through time periods on the lower section of the timeline. It makes finding articles in relevant time periods easy and fun. In addition, Time Explorer clearly features a stronger traditional search engine than Dipity. However, it remains to be seen how well this search engine will function as they expand their database. As they expand to include other publications besides the New York Times they may have to find a way to standardize how each publication ranks the popularity of their articles. In addition, since Time Explorer seems to be focused primarily on allowing users to search their database instead of actually contributing to it, someone is going to have to enter all that data. I would hate to be their intern.

Searching for topics on Dipity can be frustrating. Users cannot search for manually created topics. Only publicly created topics made through web search are part of Dipity’s search database. However, Dipity has a distinct advantage in the fact it provides a new type of search service, one that can pull in multiple types of information from multiple places. Essentially, the potential for creating content using search on Dipity is endless.
Victor: Push. Time Explorer search is as powerful as Dipity search is innovative.
News Aggregation
As we said above, Time Explorer offers a limited database of articles to choose from. At this point, users can only search New York Times articles. I am sure they plan to add other publications, in which case, the ability for users to predict future trends should only increase. I really enjoy the simplicity of the search and the ability to emphasize, add or take away keywords to narrow down exactly what they are looking for. That said, the main element of aggregation in Time Explorer is on the future search end. Users can find predictions for the future about their search topics based on past articles and trends. As the database grows, this information can only become more relevant and detailed. Very cool.

Dipity, on the other hand, focuses on aggregating news in the present tense. Most aggregated sources are based around the 20-100 most recent posts and then regularly updated. That means, the longer the topic exists, the more information will be pushed to the timeline and the more detailed a topic will become. Also, the variety of information such as videos, photos, RSS, tweets and more create a rich multimedia experience.

Time Explorer really can help predict the future, and that is not something Dipity aims to or is capable of doing. However, Dipity can create a much more rich data visualization that integrates and updates present information on current events. The two essentially explore important news topics on two different levels, Time Explorer focuses on the past and the future, while Dipity focuses on the present.
I can confidently say that people who read their news on the internet are much more concerned with current events, and not the past or future. Therefore, while Time Explorer is an effective research and prediction tool, Dipity is more pragmatic for those simply trying to get the latest update on a trending topic.
Victor: Dipity. Not saying the past or future are any less important, but these days people tend to live in the present.

Serendipity Factor
The potential for serendipity with both Time Explorer and Dipity is enormous and exciting. The two timeline tools offer an innovative and dynamic new way to search for and visualize information. While Time Explorer allows users to explore the past and how it might affect future events, they can only do so via existing content. Dipity, on the other hand, allows users to generate their own content in order to explore a topic. Dipity might not allow users to automatically reach as far back in time, but it is extremely effective at aggregating and exploring information as it happens.
Victor: Push. Both sites can lead to some amazing discoveries.
In the end, the two timeline sites offer significantly different sets of tools for data visualization. Although the overall aim diverges, many elements of what Dipity and Time Explorer intend to accomplish overlap. Better said, while Time Explorer plans to mine the future, Dipity is currently set up to mine the present. We look forward to seeing how Time Explorer develops and are excited about the possibilities.
So, our question is: what do the readers think? Do you prefer the data visualization capabilities of Dipity or Time Explorer? Would you use one or the other or both? Give both Time Explorer and Dipity a try and let us know which you prefer. If you want to share your results, feel free to do so in the comments section.
No comments
