Five undergrad students form a team in a hackathon. They quickly find out that prize money would be awarded for exploring data visualization, machine learning, and app development, as well as prize money for the best pitch presentation. Seeing that college debt is at an all-time high in the current economy, it was fear that drove these students to make an app, not inspiration.
What it does
AppHelper is a mobile app/widget that predicts user behavior, by accurately suggesting other installed apps to suit the user's immediate situation. It utilizes machine learning to analyze the current conditions of the phone, such as location, time, or battery percentage, and compare them to previous conditions when the app was used. AppHelper constantly calculates, associates, and stores the current conditions of the phone with respect to an app used, so that AppHelper can continue to update its inferences about user behavior. For users with hundreds of apps on their phone, AppHelper will save them snippets of time. A second saved is a second earned!
When the user base for the mobile app grows to a significant size, we plan to provide app analytics condensed as data visualization to app developers and mobile carriers. We really hope that our app analytics can give insight to the current market. For instance, an app developer can decide to invest in a television commercial, realizing that many people watch television in their homes, where their phones can be charged - a good audience for an app that consumes a lot of battery. Another example could be that a mobile carrier taps into a new market by building a data coverage tower in a location many people have tried to connect to the internet. In a way, AppHelper is much like Hamburger Helper - just as Hamburger Helper could also help chicken, AppHelper can really help businesses of all kinds!
Challenges I ran into
The greatest challenge was dealing with the large datasets, especially as how to make it more manageable.
And Smash Bros. was a big challenge too.
Accomplishments that I'm proud of
We're simply proud of what we did in such a small time. Once we were forced to work in a practical setting, we learned that we were much capable than we though to be.
What I learned
We all got a small taste of data visualization, machine learning, and app development. Most importantly, we were able to bounce information off of each other, and we learned most from interacting with each other.
What's next for AppHelper
The next step for AppHelper is to make these ideas even more concrete. A hackathon is a great place for brainstorming and fleshing out ideas, but there certainly isn't enough time to make a fully-marketable app. (Okay, maybe it's enough time for fiends, because just below this form is a form for a URL for a demo site or an app store listing)