facebook-pixel

Channel recommendation engine

Industry Media and Advertising, Telecommunications

Specialization Or Business Function Customer Analytics (Market Segmentation and Targeting, Upsell Analysis, Recommendation Systems & Cross Sell Analysis, Product Feature Prioritization), Consumer Experience (Customer Behavior Analysis)

Technical Function Analytics (Predictive Modeling, Data Mining)

Technology & Tools

WORK IN PROGRESS

Project Description

We have a dataset of cable TV viewers.  For each viewer we have a zip code, a list of channels watched, and average number of hours each channel is watched.  We also have a master list of channels, so that we can see what channels each viewer does not watch.

We are looking to create a channel recommendation engine.  When we get a new record (a new viewer, with their zip code and the average hours they watch each channel) we'd like to recommend which channels, of the channels they do not watch, they would most likely want to watch based on others with similar viewing habits.  The engine should return these recommended channels in descending order of likelihood to watch, with some metric of likelihood to watch.  We can also use the zip code, and demographic info that is publicly available by zip code, as additional predictors to refine the model.

The engine should be created with open source tools (R, Python, Mahout, whatever) and provided as the deliverable for this project.

We will start this project in the first quarter of 2015.

Project Overview

  • Posted
    January 22, 2015
  • Planned Start
    March 15, 2015
  • Delivery Date
    March 31, 2015
  • Preferred Location
    From anywhere

Client Overview


EXPERTISE REQUIRED

Matching Providers