MachineHack is back with an exciting challenge. We welcome all data science and Machine Learning enthusiasts to the Weekend Hackathon 16 where participants will get a chance to use existing skills and knowledge to learn more and compete with the real players in the game.

Overview

The smartphone revolution started less than 2 decades ago with the evolution of technologies like touch screen and advanced micro chips that could outmatch the processing capacities of  even some of the Computers at that time. 

Android as a handset operating system played a major part in what we call the smartphone revolution with more than 2 billion active users a month worldwide. What makes Android popular is not just its aesthetics or performance but the millions of applications or softwares that are available for the platform free of cost or paid, enterprise or consumer.

Android Application development is a demanding skill and every business has found the necessity to establish its presence in the market by having its own application made available to users.

In this hackathon, you as a data scientist must foresee the success of an android application by building an ML model that can predict the popularity of an application based on certain factors.

Data Description:

• Train.csv – 16516 rows x 11 columns
• Test.csv – 24776 rows x 10 columns
• Sample Submission – Acceptable submission format

Attributes Description:

1. Offered_By : The publisher/Organization/Company that develops the application
2. Category : The category/Genre of the application
3. Rating: The total ratings received from consumers
4. Reviews: The total reviews received from consumers
5. Size: The size of the application with unit
6. Price: The total price of the application or cost of the in-app purchases
7. Content_Rating: The content rating for the application
8. Last_Updated_On: The date at which the application was last updated
9. Release_Version: The version of the application that is currently being served
10. OS_Version_Required: The minimum Android OS version required to run the application
11. Downloads: The approximated range of downloads for the application

The participants must classify the range  at which the application falls in , in terms of the number of downloads.

Preview:

Rules

• One account per participant/team. Submissions from multiple accounts will lead to disqualification.
• The submission limit for the hackathon is 10 per day after which the submissions will not be accepted.
• All registered participants are eligible to compete in the hackathon.
• We ask that you respect the spirit of the competition and do not cheat.
• Use of any external dataset is prohibited and doing so will lead to disqualification.
• This hackathon will expire on 17th Aug, Monday at 7 am IST

Evaluation

• The submission will be evaluated using the log_loss metric. One can use sklearn’s log_loss
• This hackathon supports private and public leaderboards
• The public leaderboard is evaluated on 30% of Test data
• The private leaderboard will be made available at the end of the hackathon which will be evaluated on 100% Test data