Macadamian Blog

Trendy: From Napkin Sketch to Fandom Ready

Ayesha Khan & Dan Lanthier

From a simple “what if,” Trendy evolved from a Twitter analytics prototype into a fully-designed, functioning consumer app available on the Windows Store.

With the never-ending search for innovation comes a constant stream of new technologies and software. While this can be overwhelming, the exchange of creative ideas, collaboration, and bringing ideas to life is what Macadamian thrives on. Macadamian’s Trendy app is a great example of this.

From a simple “what if,” Trendy evolved from a Twitter analytics prototype specific to March Madness into a fully-designed and functioning consumer app available to download on the Windows Store. This side project became a platform where Macadamian’s development and UX teams were able to test their skills while experimenting with the latest technologies and software.

From the Prototype: NCAA Twitter Data Analytics

With the intent of trying out and showcasing Windows Azure, we saw an opportunity to prototype a mobile app for basketball fans that followed the NCAA March Madness Tournament. As a game approached, Twitter activity would spike with hashtags around teams, players, and all the March Madness events. This meant lots of data being collected which could be stored, analyzed, and manipulated with the use of Azure. So, we put together a quick prototype application called Mac Madness that demonstrated the ebb and flow of various hashtags associated with March Madness.

NCAA App 1 NCAA App 2 NCAA App 3

To a Product: Twitter Data Analytics

By the time we had a chance to take a second look, it was the end of March Madness 2014 and the Connecticut Huskies had been crowned as champions. Not wanting to abandon this project, we refocused the app’s concept to something more universal. Inspired by Twitter’s list of current trends, we applied the concepts introduced from March Madness to create histograms of Twitter’s top trends to visualize the data to easily compare, analyze, and discover new insights.

However, with the much larger scope of tweets came a set of challenges that required a more flexible data analytic and better optimization for computation to ensure that the data is analyzed in a responsive manner. The most notable challenges were the following:

  • Trends last for varying duration’s. (Some can last days while some can last less than an hour.)
  • Trends now have a greater number of tweets at all times, whereas trend variation and number of tweets tracked were fewer with March Madness.

Trendy 1 Trendy 2 Trendy 3

Cloud Architecture

Azure Mobile Services uses node.js which allows the use of a language familiar to most developers: JavaScript. Azure Mobile Services also supports Twitter API authentication and allows the programming of scheduled tasks that execute JavaScript code at specific intervals. This allowed scheduling queries to Twitter’s RESTful APIs: trends/place to collect the top trends, and the search/tweets API to collect tweets associated with each trend. Other scheduled tasks responsible for computing raw twitter information into graph data were also added to complete the cloud logic.

Moving most of the Twitter interaction to Azure Mobile Services offered delegation of responsibility in the application’s architecture. The cloud’s responsibility is to periodically mine Twitter for data and analyze it. Azure Mobile Services seamlessly collects the data, crunches numbers, and formats the data into something presentable for a histogram.

With this architecture, the client’s sole responsibility is boiled down to presenting that data in a creative way. The end result offers improved experience for users:

  • The client does not have to constantly poll twitter for data in the background.
  • The client does not have to do any data analysis.
  • Both imply saving a mobile device’s battery life.
  • The client does not need a Twitter account to compute this data.

What’s Next?

Trendy already provides a great proof-of-concept for Twitter data analytics. Who knew One Direction would trend so much! Moving forward, we plan to inject the application with more context. Which tweet was retweeted hundreds of thousands of times today? Why was it retweeted so much? What’s the story behind this trend? Where and when did it start trending?

The application is currently available for Windows Phone devices and is being ported to other platforms including Android and iOS using Xamarin’s development tools. This allows code sharing between the platforms.

You can download today the current version of Trendy from the Windows Phone Store.

Author Overview

Ayesha Khan

Ayesha is a Director of Engineering who leads a team of developers, QA specialists, and project managers and helps them to be exceptional at what they do. Ayesha believes that to empower her team means investing in each and every team member. Ayesha’s career in software has focussed on delivering mobile and web solutions that put the end user first. She thrives on bringing to life disruptive and cutting edge concepts. Some highlights include an industry disrupting dual screen Android phone, electronic health record applications that gave on-the-go providers access, and an instant messaging application used by millions. Ayesha holds a BA in Electrical Engineering with high distinction from Carleton University. An idea person and maker at heart you can be sure to find her working on something interesting at any given time.
Dan Lanthier

As one of Macadamian’s Development Managers, Daniel is an integral part of the software development team. Daniel has a solid background in software engineering. His expertise includes .NET applications, Azure Mobile Services, iOS, and Internet of Things (IoT) with Electric Imp and Particle solutions. When he’s not in the office managing his team, prototyping UI, or building mobile applications, you can find him on the slopes, enjoying downhill skiing.