2016 Data Science Hackathon

Preparing for the ‘Sexiest Job of the 21st Century’

According to an article from Harvard Business Review, data scientist might be the ‘sexiest job of the 21st century,’ with many companies eager to hire data science talent. Most students receive a great amount of training in data mining, visualization, and machine learning techniques through university classes. However, they rarely get exposure to real-world data sets, which are massive and complex. The goal of the Data Science Hackathon is to help students get more familiar with complex data sets during a friendly competition.

Students participated in teams of 3 to 5 people with the winning teams receiving $500 awards in the following categories: 

  • Best Machine Learning Award. Example: Utilize open source, or personally created, machine learning algorithms, to mine the mobile data to build insights, predict user behavior, and more.
  • Best Data Visualization Award. Example: Build geographic network heat maps (i.e. signal strength), or other visualization tools, to help consumers or wireless carriers gain new insights on users, devices, and/or network behaviors.
  • Best App Development Award. Example: Build Android app(s), iOS app(s), providing customers insights on their apps, device, and/or carrier networks.
  • Blackstone LaunchPad Best Startup Award: Awarded to the most viable startup team with the best pitch presentation.

The teams were judged based on three criteria: (1) the level of technical innovation and novelty, (2) the impact on the possible spread of proposal, (3) the practicality, applicability, and reliability of the proposal.

Preparation & Support

The participants had access to several resources before and during the event. Prior to the event, there were two preparatory workshops offered. These workshops provided an introduction to the applications of data science for commercial data, and were taught by the chief data scientist from Payoff.com. During the event, Data Science Mentors worked in shifts to offer consultations to the teams. 

Data

Students got access to mobile phone data from M2Catalyst, a startup that specializes in data from Android phones. The data set was more than 100GB (16GB when compressed), and contained fine-grained reports from 3,800 devices in the city of New York, for a duration of three months (from September to November of 2015). The data points can be divided into three categories:

  • App usage:
    • Meta-data on what apps were running on the devices, and their running, as well as information about resource consumption (e.g. how much memory or processing power various apps were using).
  • Battery:
    • Battery reports at various times.
  • WiFi and Network Signal:
    • WiFi and Network signal at various locations.
    • Spatiotemporal information for Network signal.
    • Signal Strength and quality.

Winners

  • Best Data Visualization Award – $500
    • Matrix Reloaded: Jianfeng Jia, Xi Zhang, Chen Li, Jiangbo Yu
  • Machine Learning Award – $500 (two winners tie, $500 each)
    • Phone Matcher: Prateek Malhotra, Kishore Narendran, Shiladitya Sen
    • Datathon2016:Tyler Ross, James Hardjadinata, Minh Thao Chan, Loïc Lesavre
  • Best App Development Award – $500
    • DataSmasherVisualization for iOS: Yaroslav Dukal, Chad Lou, David Gogokhiya, Sara Du
  • Blackstone LaunchPad Best Startup Award – $500
    • GeoMob: Andrew Wang, Chen Li, Nathan Wu, Freddie Gibson

Judging

Judging and project submissions were done using a free hackathon management service provided by Devpost.com. Teams were given a deadline for entry submissions, where they submitted their team name, team members, project code, screenshots, and a brief description. Visit http://ucidatahackathon.devpost.com/submissions to view each submission, 17 total.

Organizers

Key organizers and founders that made this event possible were: Brandon Hastings, ICS Student Council President (the driving force), Blerim Cici, Data Hackathon Founder (technical/data preparation), Homer Strong, DSI Organizer (organizer of data science mentors), Hadar Ziv, Faculty Advisor to Brandon (industry liaison to M2Catalyst partnership). Volunteers and Support from ICS Student Council, HACK Club, UCI Data Science Initiative, Bren School of ICS, ICS Informatics Department, and Employees of M2Catalyst.

Data Science Hackathon

Flickr Album Gallery Powered By: Weblizar

 

2017-11-09T10:41:35-08:00