Netflix: The first year

An evaluation of my viewing activity on Netflix

I wondered how much time I have spent watching movies and TV shows on Netflix. Not so much I thought. I don't have the habit to binge-watch anything, and I watch very few TV shows in general. For movies, I primarily use Apple iTunes and physical Blu-ray.

Netflix keeps most data hidden from the user. Upon examination of the viewing activity one may find JSON hidden in the source code. I extracted the JSON and did some super boring cleanup and less boring calculations. This is how I have spent my first year on Netflix.

The very first stream was the first episode of Fargo on Sunday, 08/11/2015, 20:34:25.597 UTC+1.

First Stream

Fargo, Season 1, Episode 1, "The Crocodile's Dilemma"

In total, Netflix counted 294,305 seconds of streaming movies and TV shows between 08/11/2015 and 31/12/2016.

Total Time

81 hours 45 minutes 05 seconds

Now let's have a deeper look at the data. Netflix counted 99,715 seconds​ of streaming movies and TV shows on Saturdays, implying that a third of my Netflix activity happens on said day. Surprisingly, I streamed more movies and TV shows on Mondays than on Sundays, but the days are almost on par. Graph: weekday distribution in minutes:

Day of week

27 hours 41 Minutes 55 seconds on Saturdays

The longest continuous streaming was 12,941 seconds of Marvel's Luke Cage, Season 1, Episodes 9-12, on Monday, 05/12/2016.

High score

3 hours 35 minutes 41 seconds on Monday, 05/12/2016

Let's analyze the time a stream was started. For the sake of simplicity I rounded all times to half-hours. Unsurprisingly, I started the majority of streams during the evening. For unknown reason I preferred starting watching movies and TV shows at full hours over half-hours. Graph: half-hourly distribution of streams started:

Time of day

22 streams started at approx. 21:00h

Analyzing the TV shows I watched the most in terms of the quickest proves to be somewhat tricky. The number of episodes and the length of the seasons vary from show to show. First of all I picked six TV shows that I watched regularly. Then I split the data into two categories: one based of the total length in minutes and another based of the number of days it took me to complete the season.

Graph: total length in minutes:

Graph: number of days to complete:

It took me 31 days to complete 543 minutes of Fargo, Season 1, while it took me 20 days to complete 535 minutes of Fargo, Season 2. Here comes the math: it equals 17 minutes per day (rounded) for Season 1 and 27 minutes per day (rounded) for Season 2. That's what I call the Speed Score. Nothing beats Bron/Broen (The Bridge), Season 3, having the Speed Score of 97 with 584 minutes watched in 6 days. Graph: Speed Score in minutes per day:

Most (quickest) watched

Bron/Broen, Season 3, in 6 days

Behind the visualization, a cleaned up data set looks like this:

"title": "Staffel 1: \"You Know My Steez\"",
"videoTitle": "You Know My Steez",
"movieID": 80002550,
"country": "PT",
"bookmark": 2780,
"duration": 2779,
"date": 1481059769659,
"deviceType": 1602,
"dateStr": "06.12.16",
"index": 226,
"topNodeId": "80002537",
"series": 80002537,
"seriesTitle": "Marvel's Luke Cage",
"seasonDescriptor": "Staffel 1",
"episodeTitle": "You Know My Steez",
"estRating": "49"

Some entries don't have any significance for evaluation purposes, others do. For globetrotter, the "country" data might be of interest as it stores the information in which country the stream has been watched, Portugal in this example. "bookmark" stores the information when the user left the stream (for continuation, in seconds), whereas "duration" stores the total length of the stream (in seconds). Bookmark exceeds duration by one second in this example, which seems impossible, but I assume it's the marker that I left the stream at the very end of the episode after the closing credits. "date" stores the information when the stream was started (in Unix time).

That's it. And is there any conclusion? Well, as my viewing activity on Netflix is concerned, one can definitely conclude that it wasn't very high between 08/11/2015 and 31/12/2016. I watched 82 hours of movies and TV shows on Netflix compared to the global average of 568 hours in 2015 (source). However, the numbers only reflect the time on Netflix and don't show time spent watching on other platforms and time spent consuming media in general.