Are Positive Titles Still the Key to Written Success?
An Analysis of Titles on Medium
So, I came up with the idea that positive titles tend to attract more attention. Analyzing titles could provide valuable insights and an interesting perspective for creating better ones.
I used a CSV file of Medium articles. Since large datasets on Kaggle (1GB or more) were too large, I downloaded a smaller file. This file contains 17,947 rows of data spanning from 2012–11–02 to 2019–05–30. The dataset includes the following variables: URL, Title, Paragraphs_num, Paragraphs_Word_Count_avg, Quotes_num, Publication, Writer, Date, Read_Time, Claps, and Responses.
The plot below shows the number of claps over time. For outlier points, I’ve added the corresponding titles because I’m interested in understanding which topics receive more claps and what they are about.
Open datasets come with some disadvantages. Since the data is accessible to everyone, many people have already explored it, uncovering interesting patterns or publishing remarkable findings. However, there’s always room for discovery. I’m curious to explore whether any new patterns still exist — particularly those related to more readable and engaging texts.
There appears to be a general upward trend in the average number of claps over time, with the most significant increase occurring between 2016 and 2017. This suggests that Medium’s popularity has grown over the years, and people are engaging more with articles.
Negative Sentiment: The average number of claps for articles with negative titles is generally lower than those with positive titles. However, there are a few spikes in the negative sentiment line, indicating that some negatively titled articles have received a significant number of claps.
Positive Sentiment: Articles with positive titles tend to receive higher average claps overall. The positive sentiment line shows more consistent growth and higher peaks compared to the negative sentiment line.
There are a few outliers where negatively titled articles have received a disproportionately high number of claps. This could be due to various factors, such as the article’s content, author’s popularity, or the specific topic being discussed.
The positive sentiment line shows more consistent growth and higher peaks, suggesting that people are more likely to clap for articles with positive titles. However, this doesn’t necessarily mean that all positive titles will lead to high claps. The content and quality of the article also play a crucial role.
Emotional Response: People might be more inclined to clap for articles that evoke positive emotions, such as joy, hope, or inspiration. Negative emotions like anger or sadness might not necessarily lead to claps, even if the content is well-written.
Attention-Grabbing Titles: Negatively worded titles can sometimes be more attention-grabbing, leading to more clicks and potential claps. However, if the content doesn’t live up to the title, it might result in fewer claps.
Author’s Reputation: The author’s reputation and follower base can also influence the number of claps an article receives, regardless of the title’s sentiment.
Limitations and Further Analysis: The plot only shows the average number of claps for each sentiment category. It doesn’t provide information about the distribution of claps within each category. It would be interesting to analyze the relationship between title sentiment and other factors like article length, topic, and author’s expertise. A more detailed analysis could involve comparing the claps received by articles with different levels of negativity or positivity in their titles.
Overall, the plot suggests that while positive titles generally lead to more claps, there are exceptions where negatively titled articles can also receive significant attention. The effectiveness of a title depends on various factors, including the content, author’s reputation, and the overall sentiment of the Medium community.
Positive and engaging titles play a significant role in capturing readers’ attention, and analyzing them can offer valuable insights for crafting better content. By exploring Medium articles through a well-organized dataset of nearly 18,000 entries, it becomes evident that patterns in title construction, claps, and audience engagement are worth investigating further. Although open datasets may have been extensively studied, the potential for uncovering new and intriguing findings — such as the relationship between readability and virality — remains vast.
Text mining, in particular, is a fascinating tool that not only allows us to dissect patterns in writing but also holds the promise of analyzing other aspects, such as emotional tone, sentiment, and even the cultural impact of written pieces. This approach opens up countless avenues for exploration in the ever-evolving world of content creation and audience behavior.
This is just the beginning of an exciting journey into understanding the dynamics of impactful writing.
To be continued…
