Today, there are about 6 million apps on the play store...apps are good and useful right? but thats only if you can find the right app based on your interests and what you are currently working on. This has inspired us to create " Personilized App Recommender " , a personilized app recommendation system , that starts off from the users web browser history..goes through three layers of filtering..and provides the user with an app that can be of most use to him based on his current interests and work(browser history) so that he can make most of his brand new smartphone.
We have been inspired by a personal experience.As regular users of coursera and regular commuters, we did now know until about two weeks that there was a coursera app that we could use while commuting.Sure,you would say that we could have just searched for a coursera app on the play store,but who has time in todays busy world and hunt for a useful app.
"Personilized App Recommender " collects your web browsing history, arranges it in order of page views , and then starts filtering.What it does is it first takes the website in history with most page visits, gets the category of the website and the tags associated with it. The first layer checks the play store for an app with the exact same name. We do this using a HTML scrapper. If found it recommends the app, or goes to the next layer .The second layer searches the play store for the category of the desired website,gets a list of 20 apps in that category , and checks for an exact same APP DEVELOPER WEBSITE as the browser history site . This ensures that the user gets the right app even if the app has a different name , but is the same as the website what the user was refering to in his browser.If found it recommends , or else goes the third layer.
The third layer goes to the play store with the category of the website , collects the first 20 apps , and get all the tags for each of the 20 apps. It then compares the tags to the tags of the website. The top two apps which have the most number of common tags as the website , are recommended.
This way the user is given an app , which will be most useful and relevant to him based on his daily activities.
We feel that this can help alot of people, apart from increasing discoverability of new upcoming apps, it gives the user a very relevant and helpful app so that he can make the most of his smartphone.The computer is basically understanding what your working on and around and it says 'hey , this app provides very useful functionality on what youve been working on ,it can probably help you, why dont you try it out'
So basically our target audience is everyone out there who uses a computer and who has a smart phone.
We are proud of the fact that it actually works and gives you the best app you should use based on your browsing history. Right now we have made a basic UI using java, but plan on improving it later . We have completed all that we planned in the alotted time with no previous preparation before the hackathon.
Thank you , The kernel