Mobile personalization using machine learning on mobile behaviour data - BC-340

Preferred Disciplines: Machine Learning, Statistics, Economics, Mathematics, Computer Science, or Physics – graduate student, post-docs
Project length: 4 to 18 months
Approx. start date: September 26th, 2017 (However flexible give or take a couple of weeks)
Location: Vancouver, BC 
No. of Positions: 1
Preferences: Language: English
Company: Mobify

About Company:

Mobify collects billions of data points each month but currently delivers an unpersonalized mobile experience. With the applicant’s data modeling expertise, we could personalize shoppers’ mobile experiences to increase the amount of money they spend with our customers. This would be innovative, not only for Mobify, but for our customers.

Mobify has a database full of raw big data. The technical challenge is developing algorithms for the data that can be tested and improved upon to personalize shoppers’ mobile experiences to increase the revenue they spend on our customers’ sites.

The applicant will first explore the current raw database and generate a working derivative dataset through cleaning, enrichment, and aggregation. The applicant will then analyze the data to develop insights and hypothesis on which user behaviours increase a shopper’s likelihood to purchase. Using these findings, the application will propose a data model and then test it on a live customer site. The applicant will then iterate based on results. If successful, we will create a generalized model that we can productize and roll out to all of our customers.

This is very innovative because real-time mobile personalization using machine learning on the type of e-commerce data we collect has never been tried. Mobify has not started an initiative around this yet.

By the end of month 3, the applicant will have a ready-to-use dataset and insights into how mobile shoppers behave. The applicant will have a hypothesis based on their analysis of the dataset. Time permitting, the applicant will have a working prototype live on a customer site and will analyze the results as they come in. The applicant will iterate on the model to improve its results. 

Background and required skills

Research Objectives/Sub-Objectives:

  • Develop algorithms for the data that can be tested and improved upon to personalize shoppers’ mobile experiences to increase the revenue they spend on customers’ sites
  • Proposed data model that is tested on live customer site – if successful, a generalized model will be created to productize and roll out to all customers

Methodology:

    • Data exploration
    • Data Cleaning
    • Data Aggregation
    • Data Analysis and Hypothesis Generation
    • Data Modelling
    • Model testing and iteration in production 

    Expertise and Skills Needed:

    • Currently enrolled in a post-doc or graduate student with concentrations in Mathematics/Computer Science/Statistics/Economics/Physics or related programs.
    • Intermediate/Advanced understanding of statistics or machine learning
    • Previous experience working on projects requiring delivered results and demos 2+ years in academic study
    • 2+ years programming experience in data processing languages (Matlab, R, Python, Fortran)

    For more info or to apply to this applied research position, please

    1. Check your eligibility and find more information about open projects.
    2. Complete this webform. You will be asked to upload your CV. Remember to indicate the title of the project(s) you are interested in and obtain your professor’s approval to proceed.
    Program: