Weekly Play Count for user_000196
Weekly Play Count for user_000196 for Lamb
Explore listening behavior by selecting a user below. You can type a country to filter by country or a number to filter by age. You can navigate to next or previous users either by clicking the respective buttons or by using the n and p keys.
After an initial exploratory analysis supported on our visualizations, we noticed that certain artists were played more frequently on some days of the week, while other artists were avoided on different weekdays.
In order to try and confirm this, we devised a simple heuristic based on the dissimilarity matrix, where we measured the "closeness" of the play count for each weekday to every other play count of the given artist. All values were normalized and combined with the inverse of the normalized weekly total play count for the user, thus giving more weight to a peak, as it is further apart from the listening behavior for the given weekday. This means that the preferred listening days for some artists might not directly map to the peaks in the weekly play count chart above!
The avoided weekdays were simply selected based on the closeness to zero. In this case, we used the interval [0, 0.1] simply to identify low values. We are more interested on the positives though. We illustrate our technique below.
The chart on the left shows the normalized weekly play count distribution for the user. As you can see, the value 1.00, on Sunday, corresponds to the day when user_000196 listens to music the most. This also means that, according to our heuristic, peaks on Sunday are less meaningful than peaks on Friday, the weekday when the user listens to music the least.