Statistical Analysis on USA Youtube Trending Videos

Visually, the number of views on each video is not normally distributed or symmetrically following a Gaussian distribution. Because the data is not normally distributed, statistical analysis that requires normal distribution assumptions cannot be used and it is necessary to look for alternative analyzes that are in accordance with the conditions of the data. The mean in each category will be calculated using the median value in each category.

It can be seen visually that the size of the data center in each category is significantly different. This indicates that each category has different characteristics which can be caused by factors such as topics, production quality, release time, and the type of content used. The top six views per category which are above the average are dominated by entertainment categories, such as Gaming, Music, Film & Animation, Comedy, Shows, and Entertainment. On the other hand, the remaining categories are below the average. Interpretations that can be drawn are:

1. The high level of user interest caused by high attractiveness.
2. It is entertainment in nature so that it has a positive side for users.
3. Larger number of users.

Based on the top 10 views videos, all of them are dominated by entertainment content. This further reinforces that content with this theme needs to be reproduced in order to attract more potential viewers so that the generated traffic can increase all the time.

Several well-known content creators are identified from the total number of views such as ibighit which is an entertainment company from South Korea which was founded in 2005. YouTube needs to give sufficient appreciation to the channel owners and carry out an efficient optimization process to recommend their content to users who have appropriate preferences. This also applies to other top channels. Other recommendations that can be made are:

1. Give awards to content creators who have good performance, such as Silver Play Button, Gold Play Button, or Diamond Play Button. This award can provide a motivational boost for content creators to continue creating quality content.
2. Help content creators improve video quality by providing guidelines for improving the quality of their videos.
3. Provide promotional support for content that is considered to have great potential so as to help that content be more easily found by potential audiences.
4. Provide suggestions for improving performance so that it attracts more audiences.
5. Providing interesting features such as live streaming, premiere features, or other new features to make it more interactive with the audience.
6. Provide funding for content that is considered to have great potential so that the content will develop and attract more viewers.
7. Create mutually supportive content creator communities such as discussion groups or online forums. This can help content creators exchange experiences and suggestions to improve the quality of their content.

Visually, the distribution does not follow a normal distribution. This indicates that there are influencing factors which on the other hand can also be a factor causing differences in the number of engagements in each category. Analysis of the number of engagement can find out how many viewers are engaged with the content. The higher the number, the more likely the content is in accordance with the interests of the audience. On the other hand, this analysis is used to measure how close the relationship between content owners and viewers is. With a good relationship, the audience will be routine and loyal to wait for every content that is launched.

The series of videos with the highest engagement is still controlled by the entertainment industry. Some of them are video clips from boyband BTS and the rest are music content from top foreign musicians. It should be emphasized that musical content really attracts the interest of the audience, which is likely to be loyal fans, plus the object popularity of the content. This high engagement indicates that the audience is actively involved in the discussion or interaction in the video.

Based on the top 10 views videos, all of them are dominated by entertainment content. This further reinforces that content with this theme needs to be reproduced in order to attract more potential viewers so that the generated traffic can increase all the time.

Quality content and potential value are based on the owner and developer factors of the video channel. YouTube needs to issue a number of recommendations in order to increase the interaction provided by loyal viewers and new viewers, including:

1. Provide facilities for direct interaction such as live streaming. Loyal viewers will be very willing to enjoy content from their idols so that the resulting traffic will be maintained.
2. Several channels with the most popular content need to be given a special space without having to block other channels. This is so that the channel owner can continue to maintain the content they produce and continue to attract new viewers who have not been reached through the recommendation system.
3. Provide analytical and statistical tools that can help channel owners to analyze the performance of their channel and content so that it becomes an evaluation material for making further content.

With these recommendations, YouTube can help content creators to increase engagement on their channels and increase audience engagement on platform.

Based on the total views of the top six categories, it can be seen that the Music and Entertainment categories experienced the most significant growth. This indicates a significant increase in user demand for content related to that category.

The number of engagement shows a very significant growth in the Music category. This industry has the potential to see the enthusiasm of users who actively participate in providing feedback. Therefore, it is highly recommended that YouTube continues to encourage and support the global music industry in order to attract new potential audiences as a top priority regardless of other videographic industries. With the growing number of views and engagement, it will be easier for the recommendations given to be widely accepted. This needs to be proven by measuring the correlation between the two things using correlation analysis techniques.

Based on the correlation analysis above, it can be concluded that the number of views has a positive and significant correlation to the number of likes and comments. Therefore, the previous statement can be considered valid.

It can be seen visually that videos with the highest engagement level have a tendency to decline. This indicates that engagement rate will saturate over time even though the number of views and engagement continues to grow. That way, before arriving at correlation analysis, we can assume that engagement rate will continue to decrease all the time trending.

Based on correlation analysis, there is a negative relationship between the number of views and engagement rate of 8.5%. That means that every significant increase in the number of views will be accompanied by a small increase in the number of engagements. This is a phenomenon that does not apply absolutely and only provides a general response from views and engagement intercorrelation.

It can be seen that the decline for each video is concentrated between zero and two percent. Even though in reality, there are a small number of videos that have a magnitude outside of this value. At first glance, the data does not form a normal distribution. Therefore, we can use the median as a measure of data center to investigate further based on its causality across all categories.

It can be seen that several categories with a very dominant number of views in the previous discussion, have drop percentage values above the average. This is quite logical considering that one of the categories, namely Music which has a significant growth in the number of views, will not always get an increase in the number of engagement in addition to views with several factors as follows:

1. The recommended content doesn't appeal to new audiences. This can be due to a lack of market research or a lack of proper content planning.
2. Poor video quality can cause viewers not to be interested in watching and interacting with videos.
3. Not so efficient recommendation algorithm.

It can be seen that the risk of declining on engagement rate is quite large for the most popular videos. Next, we will analyze the 10 videos with the most engagements.

It can be seen that the videos with the most engagements have also experienced a significant decline so that they are well above the average. We can conclude this as a certainty that no video can last forever on the trending list because new videos will always appear and competition is getting tougher in the creative industry. Therefore, YouTube as a service provider must continue to develop the most efficient algorithm in recommending a new video with a certain priority scale.

Visually, it can be seen that the data is concentrated between 1 to 3 days. Even so, there are a few that take a long time to trend so that the data is not normally distributed. Next, we will use the median to measure the data center point in each category.

In all YouTube service categories, it can be seen that the data center point is above the average. This explains that the varying amount of time required for a video to enter the trending list and is not a guarantee that a video that has been uploaded for a long time will not be trending which may be due to time and events. Sometimes there is a video that is just being discussed when the incident related to the video occurs at a different time.

Even so, the determining pattern for a video to enter the trending list can be found clearly based on the number of views that are most dominant. There is a tendency for videos in certain categories, such as Music, to trend faster because there is already a large market of users or in other words loyal viewers.

Then for videos with the most dominant interests it will also be easier to enter the trending list in a relatively short time. Based on these findings, it is necessary to have sustainable viewers who are loyal and consistent in consuming video content presented by content creators. Maintaining sustainable viewers is very important because they can help increase engagement rate and views of videos. Apart from that, they can also help increase brand awareness and create a loyal community.

It can be seen that the duration for each video tends not to be normally distributed with concentration levels ranging from 4 to 8 days.

Visually, it can be seen that the data center points for duration in each category are mostly around the mean value. Even so, there is an exception in the Music category, which is 13.5 days. This is based on the interest factor and segmentation of users who are loyal to enjoying content based on that category.

It can also be seen that several videos with the most dominant views have a longer duration than the general average. This is because videos with a large number of views have the potential to continue to get engagement and attract interest from new potential viewers so traffic will be maintained.

Furthermore, for several videos with the most dominant engagement, they get a portion of duration around the average with the exception of a handful of videos which are caused by external factors which may be in the form of videographic quality or the theme of the content contained. Release time is also one of the reasons why the video can stay on the trending list for a long time.

The visualization above provides a general description of the duration density of a video based on each quality category. It can be seen that the difference is not too significant for videos with quality from the best to the worst. Videos with excellent and good quality tend to have a longer trending duration than videos with fair and poor quality.

It can be seen that there is no significant correlation between the quality of a video and trending duration so that it is not the main factor for categorizing quality based on engagement rate because it has a blurry boundary to be recognized.

End-product quality of a video is dominated by videos with good quality with a total share of around 57% and the rest is occupied by fair and poor quality videos with very few portions of excellent quality videos. Therefore, it is very difficult to use a nominal reference for engagement rate where the value tends to decrease over time due to the decline factor experienced by each video.

It can be seen that most of the videos have a concentration of negative sentiment in the low range, which is between 2 and 10%. However, there are a small number of videos with quite extreme negative sentiments and we will find out their characteristics. It is known that the data is not normally distributed and this is a natural thing considering that the video categories are very diverse and each has different characteristics, both in terms of content and the users themselves.

Videos with the theme News & Politics have a high level of negative sentiment because this topic is often related to complex and controversial issues in society. This can trigger debate and conflict between groups with different views, and in the end can trigger a negative response from the content published.

It can be seen that the 10 videos with the top negative sentiments are occupied by videos with the theme of news from an event that smells of social politics. The value is quite extreme where the number of dislikes responses can be many times the number of likes you get. This is an evaluation material for content creators to avoid things that are sensitive and offensive in the content they produce.

It can be seen at a glance that the distribution of the data forms a uniform pattern in which it is almost certain that the data is not normally distributed. This can be used as a reference in calculating the distribution in each video category using the median value.

It can be seen at a glance that the distribution of the data forms a uniform pattern in which it is almost certain that the data is not normally distributed. This can be used as a reference in calculating the distribution in each video category using the median value.

In some videos with the most dominant number of views, most of them embed hashtags in their videos which are generally used in other videos as well. Even though the results are not significant enough, the use of hashtags remains a factor that content creators can consider as a boost to the number of views and engagement in addition to other factors, such as categories, content quality, and the number of user segments with preferences according to that context.

Some videos with a fairly dominant number of engagement are felt to have a significant correlation with common tags considering that the more and more familiar contexts used the easier the video is to find so as to attract users' interest in enjoying the content and providing a response in the form of likes and comments.

Most of the trending videos have a fairly high degree of similarity to other videos in terms of the context or hashtags used. Although not a major factor, at least using the right hashtags can help increase the number of views and engagement of a video and is a requirement for the YouTube algorithm in determining the priority of videos that will appear on the trending list.