ico-magnifying_glass
    ico-arrow-big-left

    Next Best Action Predictor - Part 2

    PRIZES

    1st

    $2,000

    2nd

    $1,000

    3rd

    $500

    4th

    $250

    Register
    Submit
    Next Deadline: Review
    1d 4h until current deadline ends
    Show Deadlinesicon-arrow-up

    Challenge Overview

    Challenge Objectives

     
    • This challenge is very similar to the previous one in the series - the difference is that you have access to winning submissions of the first challenge. The goal is to aggregate and improve them and include cross-validation to verify the results.

    • Identify the features (attributes) and trends that correlate to Financial Analysts (FAs) bringing Net New Money into the company

    • Provide a ranking of attribute sets and impact

      1. For example the most important attribute set is columns X,Y, Z as those FAs produce the following NNM.

      2. The next set of attributes is columns AA, BB,  AZ, BZ (with a justification)

      3. The the following columns have no impact (DD, DE) (with a justification)

    • Provide recommendations for what features low performing FAs should focus on, in priority order, again with a justification (see below for examples)

     

    .

    Project Background

    Our client, a global  wealth management bank, is  looking to analyze performance of their Financial Analysts (FAs) to determine the features that strongly correlate with increase in amount of managed wealth or “New Net Money” (NNM).

    Technology Stack

     
    • Data Analysis and prediction algorithms should be implemented using Python. If you want to use a different technology, ask for confirmation in the forums.

     

    Data description

    Financial Analyst (FA) performance data is provided in a single Excel sheet with about 120 features for each analyst. Features belong to one of these categories:

    • FA Details

      • Column 1 is the FA ID

    • Branch Detail

    • Product Mix (Rev)

    • Product Mix (Assets)

    • Product Penetration

    • Client Acquisition Details

    • Net New Money Details

    • Marketing Details

    • Wealth/Financial Planning

    • Client Retention

    • Digital Tools

    • Banking and Lending

    • Client Age Distribution

    • Segment Index Score

    Columns “BV-CF” identify the Net New Money. Data set is available in the forums. “Attribute Description” sheet contains description for each of the features.

     

    NOTE: New net money section has info for past 4 years, but the other data doesn’t have such details - ex Product Mix is given for the current year.  The previous years new net money might show a trend leading up to the current year, or it might not. That is for the algorithm to decide.

        

     

    Analysis requirements

     

    Main goal is to analyze the data and determine

    1. Features that are strongly correlated with bringing new net money into the company

    2. Provide a listing of the Features and Feature sets, and their impact / justification

      1. Size of the list is up to you.

    3. Determine the attributes that a given financial analyst should focus on in order for that FA to perform better (more NNM).

     

    For example, we are looking for the following output of the feature analysis (this is just an example, these statements might not even be correct):

    • Analysts having Client Retention score over 8 and have 50% of the clients younger than 45 years have 20% higher new net money than other analysts

    • Analysts with 3 year cumulative QNR over 15, are active on social network and have more than 50 newsfeed posts have 10% higher 3 year QNR and therefore 30% higher new net money

    • The most important attribute set is columns X,Y, Z as those FAs produce the following NNM.  The next set of attributes is columns AA, BB, AZ, BZ, as those with the following values (XYZ) tend to product 5.53% more NNM, than without.

    • If we have a FA with QNR=20 and isn’t active on social network, then the recommendation could be to start posting on the newsfeed (we’re looking for more interesting recommendations than this)


    It is up to you to figure out these features and a way to get there - be creative! All the features should be clearly derived from the input data set, without using any other external data sources. All characteristics should be backed up by data analysis done in Python.

     

    First 4000 data rows should be included in the analysis - for the remaining 187 rows your algorithm should output recommendations for each of the analysts.

     

    Review will be highly subjective and done by the client. No appeals will be allowed.

     

    Your submission should contain:

    • Data analysis scripts with environment setup and instructions on how to run the analysis

    • List of features correlated with bringing new net money

    • Summary document explaining the data characteristics and outlining the main findings - graphs and other visuals are highly encouraged

    • List of recommendations for analysts in rows 4001- 4186



     

    Final Submission Guidelines

    Your submission should contain:

    • Data analysis scripts with environment setup and instructions on how to run the analysis

    • List of features correlated with bringing new net money

    • Summary document explaining the data characteristics and outlining the main findings - graphs and other visuals are highly encouraged

    • List of recommendations for analysts in rows 4001- 4186

    •  

    Reliability Rating and Bonus

    For challenges that have a reliability bonus, the bonus depends on the reliability rating at the moment of registration for that project. A participant with no previous projects is considered to have no reliability rating, and therefore gets no bonus. Reliability bonus does not apply to Digital Run winnings. Since reliability rating is based on the past 15 projects, it can only have 15 discrete values.
    Read more.

    REVIEW STYLE:

    Final Review:

    Community Review Board
    ?

    Approval:

    User Sign-Off
    ?

    CHALLENGE LINKS:

    Review Scorecard