Wells Fargo
|
Next Challenge: Calling all US solvers ages 18-30, how do you shop for auto insurance online? |
Entrants must be currently enrolled in one of a select higher education degree programs in the U.S. See the attached Official Rules document for full eligibility requirements.
At Wells Fargo, our data scientists play a key role in driving innovative and meaningful insights that enable our lines of business to provide a world-class experience to our stakeholders. The Campus Analytics Challenge 2020 puts you in the role of a data scientist and calls you to develop a machine learning model. The dataset is small enough that you should be able to work with it on a standard laptop.
To help get your creative juices flowing, we encourage you to explore machine learning research literature and beyond, as you may find a creative approach in other sub-fields of machine learning.
Eligibility: Entrants must be currently enrolled in a higher education degree program at one of the U.S universities below:
- Georgia Tech
- Duke
- College of Charleston
- Arizona State University
- University of Illinois
- NYU
- U.C. Berkeley
- Stanford
- University of Minnesota
- Drake
You may be enrolled in any type of degree program including associate, bachelor's, master's, and doctoral. Please be sure to use your university's .edu email address to sign into MindSumo and complete the challenge.
- Download the Challenge Data Set.
- Download the Evaluation Data Set.
- Download and read the attached Wells Fargo Official Rules document. This contains critical information including the Key Deliverables, Challenge Instructions and Suggestions, Judging Criteria, and Winner Eligibility.
Overview:
As noted in the brief above, the Challenge is for you to develop a machine learning model to predict a binary target variable. Your machine learning model must meet: (a) the Challenge Objectives, (b) follow the Challenge Instructions and Requirements, and (c) incorporate the Key Deliverables described in detail in the Official Rules Document.
Challenge Objectives:
The key objective of this challenge is to build a machine learning model that handles
- the given data set,
- the most suitable encoding schemes,
- the least set of feature variables, and
- no correlated variables in the set of predictors
Instructions and Requirements:
When creating your Solution, you may use a novel combination of existing machine learning and/or statistical methods, or develop your own novel method in order to extract and/or represent thematic information from the data file.
The output needs to include prediction of a target variable. Additionally, your Solution must meet the following requirements:
- You must use Python 3.
- You must provide citations and sources.
Key Deliverables:
There are 3 Key Deliverables for this Challenge. Section 6b of the Official Rules document will walk you through each of the deliverables.
Please note: The descriptions for individual features in the data set have not been provided by design. As specified in rule 6a, the goal is to predict a target variable. The “training” set will have all the features + target variable. The “test” set will have all the features but no target variable. You will build the model using the “training” set. You will use your model to predict the target on the observations in the “test” set.
For any questions not covered in the attached Official Rules document, please email molly@mindsumo.