Versions Compared

Old Version 4

changes.mady.by.user Trish McCall

Saved on

compared with

New Version Current

changes.mady.by.user Lorraine Chapman

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Reconciliation, Fraud, Compliance & Analytics as a service is fast becoming a critical function in the Fin Techindustry. While many have previously tried to build homegrown solutions, few have been successful for various reasons. The DataSeers appliance provides a unique solution of accomplishing this. In the future, we would want DataSeers customers to not only know “what happened”, we also want them to know “if it will happen” or “when will it happen”. As part of the Hackathon experience, you will be provided with anonymized datasets of real transactional data, from which you will have to derive predictions around customer behavior.

Video about this challenge coming soon
Date of the eventSeptember 12 - 15, 2018
LocationMarietta Campus, J/Atrium Building
CostFree
EligibilityThis event is open to all KSU CCSE students (ACS, BASIT, IT, CS, SWE, and CGDD) who have passed their first few programming courses. Graduate students who have exempted all transitional courses or have passed at least three 5000 transition courses are also eligible.
Registration More information about the event
HPCC Systems Hackathon Team

Machine Learning for Big Data Analytics on HPCC Systems

Are you interested in how analytics can identify trends which help businesses from a wide range of markets improve their decision making?

Join our team and learn how to use the open source HPCC Systems platform and ECL-ML machine learning libraries, to build predictive models in the fields of insurance, health care, finance, security and other vertical markets.

Challenge

financial industry.

Challenge

You will be presented with two problems, you can choose either of the one. Both are real world problems and you will be working with real dataset.

Problem 1

You have been given a dataset of users on a banking platform. This dataset contains about a 100 "bad actors" a.k.k individuals on an ofac sanctions list. Given the entire data set of data and a full ofac list you would have to write an algorithm to detect the bad actors in the main dataset.

Problem 2

Merchant name cleaning and grouping is fairly common challenge in payments industry. You have been given a transactional dataset which contains transactions performed at various merchants. Given dataset contains inconsistent format of merchant names. For example, ‘walmart’, ‘wal-mart’, wal mart etc. However, all these merchants are really the same. We think using a combination of data cleaning technique and machine learning algorithm you will be required to cluster the merchant name together in the best possible way. To make your machine learning algorithm more efficient you could use the other fields provided to besides the merchant name.


Data SourceTBA
General Instructions
  1. Please follow all the guidelines specified by the data provider.
  2. Provide a brief design/tech document describing the proposed solution (2-3 pages max).
  3. Provide the bio of all participants and their roles on the project.
  4. Use the HPCC Systems platform for executing the solution.
Rating Criteria
  1. Innovation involved
  2. Understanding of Big Data patterns and its application to HPCC Systems.
  3. Team work and execution
  4. Presentation quality
Mentors available during the Hackathon

The following mentors will be on site for the duration of the event: Dan Camper, Arjuna Chala and Richard Taylor, as well as our community users from DataSeers, Adwait Joshi and Gurjot Bandasha.

The following mentors will be available to give assistance remotely: Roger Dev.More information about our mentors and how to contact them

/wiki/spaces/hpcc/pages/23579358

Slack Channelhttps://ksuccsehackathonjoin.slack.com/t/ksuccsehackathon/shared_invite/enQtNDI4MzUwMjExNzk5LTY5ODQ0ZTNjNzliZmMzYmVjZGM3ZGE3MTNkZWFkZWVkZDNhZTE1ODUyZDkxNDg3ZWM0MjdkM2NmOGNjZWMwNmM

What can I do to prepare?

...

  • Watch a quick overview video about HPCC Systems
  • Download the HPCC Systems VM. Select the operating systems you are using first and then check the VM download. Follow the installation guide instructions.
    • Note:  We will be using a cloud-based HPCC Systems cluster for the hackathon.
  • You can use your preferred editor to write queries code but we do have our own, a Windows-based ECL IDE which you can download. On the download page, select Gold and under Operating System, select Windows. Download both the ECL IDE and Client Tools.
  • VS Code is a good code editor if you don't use Windows.  Installation is slightly more complicated:
    1. Download and install VS Code from here if you don't already have it installed.
    2. Download the HPCC Systems Client Tools from here.
      1. Choose your operating system from the popup list.
      2. Choose the appropriate "Client Tools" option for your operating system.  Make sure only one checkbox on the entire page is selected.
      3. Download and install.
    3. Launch VS Code, then search for and install the extension named "ECL (Enterprise Control Language) support for Visual Studio Code".
  • Once you’re up and running, try out a few examples from the installation guide and tutorials.
  • Learn some ECL. This is the language used to write queries. It's easy to use, try it for yourself. Read the documentation or take a training course.
  • Take a look at some video tutorials
  • Take a look at the information and training examples in this GitHub repository. In particular, please look at the Taxi_Tutorial where you will find the DataSeers contribution which provides examples showing some basic ECL functionality in action.
  • Take a look at our Machine Learning Documentation and Sources.

...