Ted Talks: What Brings More Views?

Posted on Nov 20, 2016

Ted's revolutionary uniqueness brings an average of 2 million views per video. It is a dream for a video-blogger. However, the video views on Ted Talks vary on different videos. The main purpose of this project is to explore the possible reasons for more views on Ted Talks.

This is a web scraping project. All the data was scraped from ted.com by Python Scrapy module. The codes for the spider file are here. The following items are scrapped from Ted websites: video total views, comment numbers, video duration, subtitle numbers, the upload date, and the video topics.

Explanatory Data Analysis

When doing the basic graphic EDA between total views and other continuous variables, including  comment numbers, video duration, subtitle numbers, and the upload date, no prominent or visible trend or associations are found. The categorical variable, topics, shows some useful information on views. Since each Ted video involves several topics,  views comparisons among different topics are in the form of one topic vs. other topics.



Compared with other topics, psychology, business, and entertainment attract more views, while videos of technology, global issues, science, and health have a relatively low view numbers. Those results are displayed through graphic EDA, and further analysis needed to prove the results are statistically significant. However, topics seem to be an important factor of video views. A topic perspective is provided in the next section.

Topic Perspective

The network here has a different meaning. That is, more connections with other topics, more often this topic is used with other topics in Ted videos. Here is a topic "network" graph. This is an interactive graph, and you could check it here.

screen-shot-2016-11-19-at-11-37-10-pm screen-shot-2016-11-19-at-11-37-36-pm

Each circle is one specific topic. Each color represents one topic type. The bigger the circle is more connections with other topics. The shorter the link is, more often this link between those two topics is. Combining aforementioned graphic EDA with the network graph, when a video involves more topics in it, the video has a relatively low views. In other words, people may be more interested in a video focusing on one field.

All the inferences above are coming from graphic observation, and it is necessary to have further statistical analysis to confirm the conclusion.

Statistical Analysis

This project, firstly, compares means of views between one topic and other topics, and no significant differences are found. Then, a multiple linear regression model is attempted here. Because the distribution of total views is skewed, a boxCox transformation is performed. The lambda value here is -0.14.

Based on stepwise regression and AIC, BIC model comparisons, the final model is as follows:


The p-value of F-test for the whole model is significantly small, which supports the validation of the model. Although the transformation of the views makes the model hard to interpret, we still can get useful information here. What the model tells us is that with more subtitles, more comments, and longer video duration, the video will gain more views; however, the older the video, referring the upload date, the fewer the views. In other words, people are more interested in newly uploaded videos. When it comes to topics, the more attractive topics are psychology, business and entertainment, which is displayed by the graphic EDA. Global issues and health topics are less preferable by people when watching a Ted talk. People show no preference for technology and science topics.

End Note

To sum up, when people to choose to watch Ted talks, what are the factors influencing people's choice? The subtitle matters, other people's comment matters, the video duration matters. and, most importantly, the topic matters. People are more interested in psychology, business and entertainment topics instead of global issues and health topics. For future improvement of the model, more data from different perspectives will help build a more perfect and interpretive model.

About Author

Related Articles

Leave a Comment

No comments found.

View Posts by Categories

Our Recent Popular Posts

View Posts by Tags

#python #trainwithnycdsa 2019 airbnb Alex Baransky alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus API Application artist aws beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep Bundles California Cancer Research capstone Career Career Day citibike clustering Coding Course Demo Course Report D3.js data Data Analyst data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization Deep Learning Demo Day Discount dplyr employer networking feature engineering Finance Financial Data Science Flask gbm Get Hired ggplot2 googleVis Hadoop higgs boson Hiring hiring partner events Hiring Partners Industry Experts Instructor Blog Instructor Interview Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter lasso regression Lead Data Scienctist Lead Data Scientist leaflet linear regression Logistic Regression machine learning Maps matplotlib Medical Research Meet the team meetup Networking neural network Neural networks New Courses nlp NYC NYC Data Science nyc data science academy NYC Open Data NYCDSA NYCDSA Alumni Online Online Bootcamp Open Data painter pandas Part-time Portfolio Development prediction Prework Programming PwC python python machine learning python scrapy python web scraping python webscraping Python Workshop R R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn Selenium sentiment analysis Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau team TensorFlow Testimonial tf-idf Top Data Science Bootcamp twitter visualization web scraping Weekend Course What to expect word cloud word2vec XGBoost yelp