Last week, Narrative Science’s Quill Divvy Bike Story application was announced as a finalist in the Most Creative category of the Divvy Data Challenge 2015. The Marketing Team sat down with Narrative Science’s Dan Platt and Craig Booth to dig a little deeper into their Divvy Bike Story app. The contest closes April 7th, and winners will be announced April 8th.
Vote for Dan and Craig today!
Q: How’d you hear about the data challenge?
A: We have presented at OpenGov Hack Night before and follow a bunch of people from that event and saw it all over Twitter, plus we knew about it from last year.
Q: What’s your app all about?
A: Our app allows a user to put in any address or zip code within the city, and we use Narrative Science’s Quill to write a story about the closest divvy station to the address the person entered, look for interesting metrics to measure that station and compare it to the rest of the stations in the city.
Q: How does it work?
A: Utilizing publicly available data from the City of Chicago, we take data from 3 million Divvy trips, categorize the trips by neighborhood and tie that information together with weather data. Then, Quill analyzes this data and determines the five most interesting things about each station by identifying the outliers — the ways in which this station is different than the others.
When a user enters their address and clicks the “Write Story” button, Quill generates a story about the Divvy station closest to that address. The narratives Quill generates are updated as the underlying data changes, so each user receives a fresh story no matter how many times they click “Write Story.”
This is the first public facing application which transforms Divvy data into easy-to-read narratives, which is really cool, and to complement the narratives with a map is cool, as well. The stories are dynamic, personalized and enhance accompanying visualizations. We know the information provided is always fresh each time you click the button, which is fun.
It also shows the power of our tool that we can integrate narratives into other visualization mediums. You’re not just hitting a button and getting a narrative, you’re clicking around a map that has pins on it and you’re getting insights specific to the area you clicked.
Q: What kinds of insights does Quill provide using the Divvy data?
A: Quill tells you what’s interesting about your station in comparison to the other stations by calculating around 30 metrics and writing about the top three to five most interesting trends.
It does this by incorporating analysis on different metrics, including the number of riders who use the station, amount of journeys from that particular station in super windy weather, location with the most morning or evening commuters and comparisons on characteristics such as riders’ ages and overall station density.
A: We felt as if the ID of the bike is something that really only Divvy would know or care about (unless you’ve found the red one), so you as a rider don’t know that you were on a certain bike. However, the actual station that’s in your neighborhood actually means something to you.
Quill could’ve just as easily written stories about bikes, routes, dates or any other way you care to slice and dice the data, but people have a personal connection to places, and we wanted Quill to write stories that were meaningful.
Q: How much time did you spend developing the app?
A: We went from nothing to the submission-ready-app in less than a month as a side project. We had stories really quickly, and then it was just thinking about bells and whistles in the app, which is all bootstrap.
Q: Why do you think your app should win for Most Creative in the Divvy Data Challenge?
A: I feel like we show the data in a truly unique way. Before we built ours, we looked around at finalists and winners from last year, and everything in the contest is very visual. People made really cool infographics, but there were very few interactive applications where people could click around, explore and get immediate insight in plain language.
The idea that you can dynamically generate your own narratives within an application is something that is unique to us in this contest. It is something that no one else does. It’s a very sophisticated way of using the data that’s available from Divvy but provided in a very easy-to-consume way.
We believe our technology can help inform the public, and there’s so much data that is being made available from the City of Chicago. This is just another display of an application you could build with our software from publically available data.
Q: What are you going to do if you win?
A: Have a beer (they both laugh). But first, everyone needs to vote, and we will gift Narrative Science the Xbox if we win. In all seriousness, though, we didn’t enter the contest for the prizes. We entered because we believe Narrative Science has a technology that helps make people smarter, and we saw a fun opportunity to take data that was made available and to build something fun.
We just like building cool things. Normally, we operate in industries like financial services and government, so the fact that we could build this app was a nice fun side project for us. This isn’t the first time we’ve collaborated on something like this, and as long as there’s publicly available data that Quill can write stories about, it won’t be the last.