by Dawnis Chow @ Alphonso Data Science
April 18, 2019
We humans are creatures of habit. It’s no surprise that these habits become evident in TV viewing data, allowing us to discern the personality of TV-watching households. Here we explore month-to-month consistency in the number of unique networks watched by viewers. We can see that households fall into distinct classes based on their behavior. In order to study the robustness of TV watching behavior over time, we tracked randomly-selected households over the course of six months from October 2018 to March 2019, sampling TV watching habits during the first 9 days of each month.
The unique networks watched by a household on a month-to-month basis is highly consistent.
Normalizing by the amount of time spent watching reveals consistent behavior in the number of unique networks watched per hour.
Household Behavior Partitions by Unique Networks Watched per Hour
These analyses suggest that we can classify households with robust behavioral metrics. It is possible that households that exhibit high unique networks watched per hour behave differently than those with low counts in other ways as well. More work remains in exploring novel metrics and their relation to consumer behaviors outside of TV watching. Watch here for more insights from Alphonso Data Science.