In San Jose, CA, a small group of designers, developers, and data scientists came together to bring design to Spark, an open-source high speed engine for big data computing. Our idea was an app called RedRock, which utilizes Twitter data to gather insights through the use of visualizations and runs on Spark. Within 10 days, we had a full design, 14 iterations of working code, and trained algorithms. The true power of RedRock, however, that is lowers the barrier to understanding the power of big data and analytics to create an approachable and non-intimidating experience for the user.
When we started this project, we had a rough prototype which was the original RedRock that came out of a development project for an IBM Hack-a-thon. As a design team, we took a hard look at the prototype and looked for opportunities to add experience and to understand the problem that we needed to solve for the user. It had the technology behind it but not the experience.
We began to sketch about what this app could become. Diverging and converging as a team, we explored the different ways that a user might interact with Twitter feeds, the different types of data that we could pull and visualize, and interactions. Here are some of the first sketches of RedRock alpha:
Once we converged on our sketches, we used Keynote as our primary wire framing tool to quickly present and iterate on our ideas for the app. For RedRock, we focused on a marketing analyst user who works with social media. They need a very non-technical tool where they can view topics they are interested in, understand the overall landscape of that topic, and dive deeper into it to find areas of insight. Understanding this, we wanted to create an iPad app that focused heavily and displaying metrics from Twitter in different lenses.
With a few more designers joining our team, we brought many iterations to the table before deciding on the final color palette and layout. We finalized on a dark blue and teal color palette because it didn't distract or overwhelm the bright colors of the data visualizations.
Supplementary designers: Andrew Smith and Voranouth Supadulya
Working hand in hand with development, the team pushed 14 iterations of the app to TestFlight in those 10 core days. When we were not all collocated in the same room, we would iterate together on Skype.
Everything came together to create the working app, RedRock. How the app works: The user enters a hashtag that they are interested in search, i.e. #sports. The app then calls out to 200 million terabytes worth of data from the Twitter decahouse and then delivers back a feed of the tweets the the top impressions that feature your searched term and visualizations. The final app includes 6 visualizations; a cluster graph of the related topics to your searched word, a sentiment bar chart, a treemap of the occupations that are tweeting your search word, a map that highlights the locations that the tweets are coming from, and a network graph of the top 20 words that appear with your searched word. The app has now been shown at many events, including the Spark Summit 2015 for a group of analysts and a main stage demo for 3,300 people in Las Vegas at the IBM Analytics Sales Academy 2015, both of which I presented. RedRock, however, is not a product but a model of what we could do in a short amount of time by bringing designers, developers, and data scientists together. Below, you can find a video of the presentation of the design process and a video of the demo of the app.
Here is a video of me presenting the design process the team undertook to create RedRock. This was presented at the Spark Summit 2015 event to a room of analysts.
In this video, Steve Beier, Program Director of the Spark Technology Center, demos the RedRock alpha app during the Spark Summit 2015 for a group of analysts.