For the final, I continued to explore the theme of loss, creating a static collection of lost objects from Twitter using the query “I lost my.” In the process, I started to tackle the Twitter API and Node.js for the first time. As a result, the technical aspects of the project took the most time and focus, but the result—a site with a collection of words and a simple interaction that reveals their stories—is a good jumping off point for this theme that I’ve been attached to: the range of everything encompassed in the word “lost.” If you click, the objects will change color according to the categories “people,” “things,” “states of mind,” and “other.”
A couple of lessons:
• I knew learned this in ICM, but rediscovered that Daniel Shiffman’s videos are amazing. His tutorial on Node.js saved the day (many days, for that matter).
• A big lesson throughout this process was that you can learn a lot by Googling things, but you have to know what to Google in the first place. Once I figured out how to collect tweets with Node, I then had to learn how to write the data from the server into a JSON file, but because I didn’t know exactly how to phrase my question, I ended up rigging a formula that would write the tweet text into JSON format with a bunch of quotation marks and commas, basically manually forcing it into the write format. When I ran the project by Jer Thorp at office hours for Data Art, he corrected the very silly error of my ways with JSON.parse() and JSON.stringify(). You don’t know what you don’t know.
Takeaways and hopes for the future:
In the future, I want to tackle the Twitter Stream API, so that the collection of objects can be dynamic and change as new tweets come in. This would add the dimension of time, which is very tied to loss in all its forms.
In its current form, the project relies on a lot of manual work. I ended up collecting the lost objects myself because I couldn’t get my formula for automating it to work (I tried separating each tweet into an array separated by spaces, then finding the word after the phrase “i lost my” but before the index of the next space). The upside to this is that I got to have more editorial say in the objects. Keeping two words in “fucking ball,” “Beats headphones,” and “abortion money” are all choices I could make. Still, more automation would be nice and even necessary if it runs on the Tweet Stream API.