This project was designed to put the enormous Google Ngrams dataset to good use. What could we do with half a trillion database records? We decided to tell a story about the words and phrases we hear every day.
We learned a lot from this experiment, and not just the meaning of the word Relevance or the extent of Shakespeare's contributions to the English language. This was an exercise in how to build apps that clearly and efficiently distill massive datasets into a human-consumable format. We met that goal.
We also learned some hard lessons about the state of various technologies and libraries dominating Big Data applications right now.
Now that we've learned what we needed to learn from Word Magic, we have shut it down. While Hadoop is great for scaling up, it doesn't do as well at scaling down.
We hope you enjoyed playing with Word Magic if you had the chance, and may one day spin it up again when there is a new generation of Big Data technologies for us to learn.