The goal of this project was to model the lifespan of news stories over time and over different topics. I extracted 17.25GB of data from the New York Times containing the first paragraph of every article they had ever published. I used that data to create files which together covered every word stem and which contained the time series of the popularity of that word stem over time. (The popularity of the word stem "war" over time is shown above as an example.) I then creates a method for characterizing the shape of an average peak in each of those time series. That gives us the average "shape" of a story about that word stem. (The bottom image shows the average shape of stories about the word stem "tsunami". There is more coverage of tsunamis after the peak than before, which makes sense because there generally isn't much coverage of a tsunami before it happens.) I calculated those shapes for each decade, averaged them, and then measured the width of each of the averages to measure the average lifespan of stories in that decade. In the end, I found that the lifespan has been getting consistently longer since the 1850s (the whole range of the data).