Steemph.cebu bot - A Community Upvote Bot Performance

INTRODUCTION

---

Steemph.cebu Bot is a steem blockchain upvote bot for Philippine Cebu Community developed The bot aims to effectively curate all Cebuano authors. This bot is link to the the Cebu community account which is @steemph.cebu. In addition, the bot is backed with 464 curation trail follower hosted at steemauto.com.

Currently, the bot's developer release two version as stipulated in the bot's Github repository TeamMalaysia/Steemph.cebu. The versions release has the following descriptions:

• Version 1 allows @steemph.cebu curation trail followers to upvote its post using $upvote <steemit link> with a 12 hours cool down on the community discord serverl Teenvestor ; and,

• Version 2 has two updates made. The admin channel with no cool down, and added adjustable vote weight. Adjustment of vote weight can only be done by the admin ( Cebu community core leaders).

The steemph.cebu curation trail listed a total followers of 464 authors which only 180 authors has been activated in the Teenvestor discord server as steemph.cebu curation trail follower as of 10:22 AM, April 28, 2017. The list was verified by the channel's moderators and admins. The server data was obtain with the help of Teenvestor moderator. An author is allowed to used the discord command for the bot when activated and designated as steemph.cebu curation trail follower in the Teenvestor discord server.

OBJECTIVE

---

The goal of the analysis is to evaluate the voting performance of the @steemph.cebu upvote bot in order to:

• to evaluate the @steemph.cebu bo performance based on the average payout, vote, and vote weight; and,
• to analyze voting behavior of the @steemph.cebu bot .

OUTLINE

---

I. Scopes of Analysis

II. Tools and Query

III. Data Presentation and Result

III-A. Number of Post receiving an upvote from @steemph.cebu bot.
III-B. Payout value of post voted by the @steemph.cebu bot
III-C. Number of Votesand Vote Weight per post voted by of @steemph.cebu bot

IV. Conclusion

I. SCOPES OF ANALYSIS

---

The analysis performed using the data extracted from SteemSQL database on 6:00 pm (UCT), April 27, 2018. The data extraction is concentrated on a 3 month time frame between January 01 to March 31, 2018. The following analysis is based on all posts which @steemph.cebu voted from the specified time frame.

In addition, the query involves in this analysis is categorize based on total_payout_value as to the following payout range:

• average payout less than 1;
• average payout between 1 to 100; and,
• average payout greater than 100.

This is done to have a good visualization of data and eliminate some outliers from the full data set. Furthermore, vote and post counts are also included specific to the different categories and authors that has received the @steemph.cebu upvote bot. Average vote count and weight is also considered in the analysis. This parameters are extracted from two tables in the steemSQL database:

• Txvotes; and,
• Comments.

Aside from looking into the steemSQL database, the steemauto.com curation trail @steemph.cebu's list of followers with the allowed voting weight is considered. This is used to co-relate the extracted data for vote weight to the allowed vote weight.

Moreover, historical commits data from the TeamMalaysia/Steemph.cebu Github repository, where the initial commit start on February 22, 2018, is considered. The first series of commits ends on February 24, 2018. The second series of update starts at April 7 and ends April 9, 2018. The first and second series of commits are for the Version 1 and 2 of the @steemph.cebu upvote bot.

Lastly, post authored and upvoted by @steemph.cebu account was excluded in the analysis.

II. TOOLS AND QUERY

---

The data extracted from the SteemSQLdatabase using Microsoft Excel. All data processing was done in Microsoft Excel which includes data visualization and statistics. The core query to find all post voted by @steemph.cebu bot:

SELECT                  
    MONTH(Comments.created) AS [MONTH],             
    DAY(Comments.created) AS [DAY],                 
    COUNT(Comments.permlink) AS [NO_OF_POST],                   
    AVG(Comments.net_votes) AS [AVERAGE_NET_VOTES],                 
    AVG(Comments.total_payout_value)  AS [AVERAGE_PAYOUT]                   
FROM Comments (NOLOCK)                  
INNER JOIN TxVotes (NOLOCK)                     
        ON TxVotes.permlink = Comments.permlink AND         
                   TxVotes.author = Comments.author                 
WHERE                   
    YEAR(Comments.created) = 2018 AND               
    TxVotes.timestamp >= '2018/01/01' AND                   
    TxVotes.timestamp < '2018/04/01' AND                    
    TxVotes.voter = 'steemph.cebu' AND                  
    TxVotes.author != 'steemph.cebu'                                    
GROUP BY                    
    MONTH(Comments.created),                
    DAY(Comments.created)                   

All votes weight from post the voted by @steemph.cebu bot can be access using the query:

SELECT      
    MONTH(TxVotes.timestamp) AS [MONTH],    
    DAY(TxVotes.timestamp) AS [DAY],        
    COUNT(TxVotes.permlink) AS [NO_OF_POST],        
    AVG(Txvotes.weight) AS [AVERAGE_VOTES_WEIGHT]       
FROM TxVotes (NOLOCK)       
INNER JOIN Comments (NOLOCK)        
        ON Comments.permlink = TxVotes.permlink AND
           Comments.author = TxVotes.author     
WHERE       
    YEAR(Comments.created) = 2018 AND   
    TxVotes.timestamp >= '2018/01/01' AND       
    TxVotes.timestamp < '2018/04/01' AND        
    TxVotes.voter = 'steemph.cebu' AND      
    TxVotes.author != 'steemph.cebu'                
GROUP BY        
    MONTH(TxVotes.timestamp),   
    DAY(TxVotes.timestamp)      

Categorization based on payout was done by adding a WHERE condition on the core query:

/*average payout less than 1 */
    Comments.total_payout_value <  '1'

/*average payout between 1 and 1000 */
    Comments.total_payout_value >=  '1' AND 
    Comments.total_payout_value <  '100'

/*average payout less than 1 */
    Comments.total_payout_value >  '100'

III. DATA PRESENTATION AND RESULT

---

III-A. Number of Post receiving an upvote from @steemph.cebu bot.

There are three highlighted section in the daily number of post voted by @steemph.cebu bot plot. This sections were divided in accordance to the movement of the daily post count as to average post count. To understand the post count trends, the @steemph.cebu bot was ready to use on February 25, 2018 where @steemph.cebu curation trail follower was able to vote post using the discord channel command $upvote <steemit link>. This only allows user to use the bot every after 12 hours cool down. Beforehand, curation trail follower as activated in an automated voting from the admin, receives the @steemph.cebu vote as well as the trail, which started on January 2018.

The total post count for the full timestamp equal 8952 over 549 unique authors. It is highlighted with a 47.78% share of time where post count is above the average 99 post, situated in the plot between timestamp Feb 1 to March 15. From February 25 to March 16, 2018, a downward trend is observed highlighted by 63 post count drop. This drop is a good indicator that there is a significant decrease in the number of spamming and cashing out post being voted by the @steemph.cebu bot.

A good result from the established 12 hour cool down is indicated in the numbers in the 17.78% timestamp share where daily post count below the average 99 counts from the plot. This lower post count average shows a better curation of quality post. This is a good indicator that there are less low quality and spamming post being voted by the @steemph.cebu bot, within a specified timestamp between March 16 to 31, 2018.

The 34.44% time share where daily post count neither concentrated below or above the average 99 counts from the plot. At this timestamp, the curation trail is yet gaining members and admins updating the voting list. The increasing post count from January 5 to February 24, 2018 shows growth in the number of curation trail follower; and also suggest an increasing number of spamming and low quality posts.

Overall, the @steemph.cebu bot voted post was trimmed down from February to March 2018. The 12 hours voting cool down affects the number of voted post per day. March post count were far low from the average of 3 months. However, the low post count has a positive effect on the post voted by the @steemph.cebu bot. It trims down spamming and low quality post which only aims cashing out upvotes from the bot and its curation trail. It is projected that low post count will be sustained in the upcoming months, as the community and developer will emphasize quality over quantity.

III-B. Payout value of post voted by the @steemph.cebu bot

Dividing the payout plot into three clusters leads to a better look into the daily average of the total payout value. The cluster made are based on the observation of the full data extracted from the SteemSQL database. The total of 8952 post count that was voted by @steemph.cebu bot where each clusters has the following post counts: 5662, 3275 and 15 for Cluster 1, 2, and 3 respectively.

Cluster 1 comprises 63% of all data points with an 0.2368 average of the average total payout value. Most total payout value averages in Cluster 1 is below the average line, which is 72.09% of the time. Two trends can be seen on the Cluster 1. A downward trend from January to early March; and, an upward trend in the until end of March. This upward trend, as shown in the end of the plot, shows a positive indication that the Cluster 1 data will have reach and sustain a daily total payout values above the current average in the upcoming months.

Cluster 2 shares 17% of all data points having an 8.77 average of the average total payout value. A downward trend in daily averages is shown in the plot. A big drop in the daily total payout value average where it hit rock-bottom at 2.06 in March from a 30.28 max point in January. A projected downward trend will possibly be sustained in the upcoming months. This continuous down in data points in Cluster 2 suggests a more occurrence of data under Cluster 1 in this upcoming months.

Cluster 3 occurrence in the next month is not anticipated. Data points in this cluster is concentrated in January with one occurrence in March. The data points from this cluster is mainly from 15 post only. Cluster 3 is just full of data outliers from the full data set. No data trends can be observe and, at the same time, no fixed time interval for the data points. Average value shown in the plot is the average total payout value against 15 post count.

The 63% possibility of occurrence of Cluster 1 suggest that average total payout value is equal to the average of Cluster 1 with the 0.085288 standard deviation. So, post voted by the @steemph.cebu bot would foresee an average total payout value equal to 0.151504 to 0.32208. On the other hand, A 17% possibility that a post would fall under Cluster 2. It is expected to have the an average total payout value amounting to 3.93 to 13.61. However, the downward trends suggest a value around the lower limit of the 8.77 average.

III-C. Number of Votesand Vote Weight per post voted by of @steemph.cebu bot

The average vote weight cast by @stemph.cebu bot is 6.827% on top of the 53% allowed average vote weight by users for the trail. The plot shows the average vote weight for February to March 2018. The January vote weight was not included to have a better look at the February to march data trends.Hence, if we look into the the bot's repository historical commits, the February is a critical point of the bot's development as reflected on the weight adjustment feature on the bot's version 2.

As stated earlier that February was a turning point for the bot and its trail, the plot shows that at the bot effective usage on Teenvestor discord server in February 25, which leads to a positive increasing trend in the vote weight average and the vote counts per post. On average, post gets an average upvote by @steemph.cebu bot and its trail equal to 105 upvotes . This is 22.63% of the total votes a post can receive from the 464 curation trail follower. In addition, it is only 38.79% of this curation trail followers where activated in the voting list.

A down trend in the plot can be observe before February 25. This was due to the the transition from the automatic post voting by which all post of authors on the voting list is automatically voted by @steemph.cebu bot. This was subjected with massive abused by authors who only want to profit the system. After the implementation of the 12 hours voting cool down, and the move of voting by the @steemph.cebu bot to the discord server, an increasing trend of the vote weight was seen, which reached the 15% mark. In terms of the number of votes cast , a stable numbers of up and down is observe in the plot, which are above the 105 average line.

IV. CONCLUSION

---

The following conclusion can be drawn from the analysis:

• The number of voted post by @steemph.cebu bot was cut down, which was highlighted by17.78% of the time frame . The downtrend in the post count is a good indication that post intended to profit the bot was trimmed down by the implemented 12 hours voting cool down per author;

• There is 63% chance that a post will have an average total payout value equal to 0.151504 to 0.32208. Another 17% possibility that a post would have an average total payout value amounting to 3.93 to 13.61. The downward trend in Cluster 2 (average total payout value between 1 to 100) suggest that Cluster 1(average total payout value less than 1) has a high chance to occur; and,

• The average vote weight cast by @stemph.cebu bot is 6.827% on top of the 53% allowed average vote weight by users for the trail. Both the voting weight and votes count per post has increase after the implementation of he 12 hour voting cool down per author. Average vote counts per post after february 25 sustained above the average 105 votes line.

---

Posted on Utopian.io - Rewarding Open Source Contributors

---