Weekly Dose of Data Analyst#10 — What is the optimal reading time for your medium posts to get more claps and responses on Towards Data Science Publication ?
Problem:
Towards Data Science is one of the most popular medium publication platform which shares information about Data Science. It is my go-to place for getting new information about AI/Data Analytics. And it is one of my goal to get published on that platform one day.
I am curious about what is the optimal reading time of a medium post to get more claps and responses. Here is my finding.
Similar Studies:
Someone has done similar studies before about optimal reading time, however their post is using time spent on reading posts as criteria. In addition, their post is outdated (This is done is 2013) and is not specific to Towards Data Science publications. If you are interested, you can click the link here to read it (The Optimal Post is 7 Minutes. Understanding which Medium posts get… | by Mike Sall | Data Lab | Medium).
TLDR:
The optimal reading time for more clap and response is 10 minutes.
Detailed Analysis:
This is based on median number of claps for each reading time, generally the longer each post’s reading time is, the more claps that post will receive. The most efficient reading time to get the most clap is 10 minutes where the median number of claps is 133 (See screenshot below).
It is more difficult to observe trend based on median number of response for each reading time, as more than half of the posts do not receive any response (See screenshot below). However, the median number of response increased from 0 to 1 when the reading time reaches 7 minutes. This suggest that if a post having a longer reading time will definitely help on getting more responses.
Preparation Process:
- Following data is required:
- Reading time of each posts published in Towards Data Science
- Their respective number of claps and number of response
All above information can be extracted from Towards Data Science’s archive website(https://towardsdatascience.com/archive).
2. Data Acquisition & Extraction:
Python code is used here to extract the number of claps and number of responses using its web scraping package.
This Python code is revised from Dorian Lazar’s web scraping code for medium. You can see his code and post in the link here (Scraping Medium with Python & Beautiful Soup | by Dorian Lazar | Medium)
Please click the link here to access the Python code.
3. Data Examination:
We can see the number of posts get published in Towards Data Science increased almost exponentially during the last 5 fives. It grow from 47 posts per year in 2016 to 2020 where 19549 posts have been published.
Looking at the frequency distribution of number of response & number of claps for all reading time. It is easy to observe that the distribution is positively skewed. This distribution does make sense since it is more likely for people to not comment and do not clap on a post. One interesting thing is that there are quite a few posts which has exactly 1000 claps and 2000 claps.
4. Data Cleaning:
Since the median & average number of claps and responses are being compared in here, the minimum sample size of 100 is used to make sure the numbers are more representative for its reading time group.
This means if the number of posts with the same reading time does not exceed 100, it will not be counted into the results. As a result, the posts with more than 19 minutes reading time has been filtered out.
All above charts can be found in this tableau workbook.
If you finish reading this post and liked it, please click that clap button and let me know your thought.
