In a surprising move, Netflix has broken its usual pattern of secrecy by publishing a public dataset that includes viewership numbers for more than 18,000 titles. The dataset covers a six-month period from January 2023 to June 2023 and represents a staggering 99% of all viewing on Netflix. According to the company’s blog post, this dataset accounts for nearly 100 billion hours viewed, showcasing the immense popularity of their content.
The Significance of “Viewership Hours”
Netflix measures viewership based on hours watched rather than individual viewers or households. This approach acknowledges the fact that individuals may watch a show or movie multiple times. It’s a more comprehensive metric that truly reflects the amount of time users spend engaged with Netflix’s content.
Exploring the Data with ChatGPT
The author of the article decided to delve deeper into the dataset by leveraging the capabilities of OpenAI’s ChatGPT. Using ChatGPT, they performed a data analysis to uncover insights. ChatGPT provided a concise and clear analysis, highlighting key findings and points. However, it encountered some challenges when asked to generate specific visualizations or charts.
“The ‘Release Date’ column has a significant number of missing values (13,359), which may limit certain types of time-based analyses.”
ChatGPT successfully provided details on the top 10 most-watched titles during the Jan-Jun 2023 period, as well as the least viewed titles in terms of viewership hours. It also determined the median viewed title and the average hours viewed. However, when asked to generate a line plot showing viewership hours on a month-by-month basis, ChatGPT struggled due to the absence of specific monthly data in the dataset.
The author explains that the plotted chart displayed cumulative viewership hours over the entire six-month period for titles released each month. It does not accurately reflect monthly viewership. The author found it challenging to guide ChatGPT in creating an appropriate and correctly labeled chart.
ChatGPT as an Analysis Partner
While ChatGPT proved to be a helpful tool for conducting data analysis, the author emphasizes that it still has room for improvement before becoming a fully reliable and insightful data analyst. Although it provided valuable insights, creating accurate visualizations required several iterations of back-and-forth interaction.
In conclusion, the Netflix engagement report’s release marks a departure from the company’s secretive approach. By embracing transparency, Netflix has given researchers and analysts access to a wealth of valuable data. ChatGPT shows promise as an analysis partner, but its limitations and the need for human guidance underscore the importance of human expertise in the data analysis process.
