Repository
https://github.com/steemit/steem
Aim of Analysis
Every month I produce a Post bench-marking report with the aims of:
• Aiding the STEEM Community with a place to benchmark their posts and growth
• Create a report that can be used for comparisons in other periods
• To track my personal targets and record achievements
Please note this data is based on Posts and not Comments.
If you missed last months report you can check it out here
/@paulag/september-18-steem-post-bench-marking-report-blockchain-business-intelligence-1539072808561
Looking at September we had HF20, which caused a major disruption in activity on the block for the last week of the month. Votes, Comments and Posts were all down. However in October Resource credits introduced in HF20 should have leveled out and values for October should reflect the true position for posts and votes and comments. However the rewards pool was inflated also due to the roll out of HF20 and so payments should be inflated in October while the rewards pool rebalances.
Financial Performance
In October 18 the % of posts with no payout was just over 16%, down from 28% last month. This is a massive change.
The % of posts with a payout of <$1 has increased from 52% last month to 56% in October.
The number of posts with a payout of over $500 remained consistent with last month on 5.
The average post pay-out value in October was $2.45 ( up from $1.6 last month) and if we exclude posts with no payout this increases to $2.94
The median payout value was $0.13 with a standard deviation of 10.58
The number of posts decreased by 13% on the previous month and the gross payouts increased by 33%
You can see how this compares to previous period in the charts below.
Top Authors based on Gross post payouts
Top Category
Comments
The % of posts with no comments is up from 36.6% in September to almost 41%
The mean number of comments per post is 3.14, and the median is 1. You can see how this compares to previous period in the chart below.
Votes
The % of posts with 0 net votes has reduced from over 11% in the previous month to 7.47% this month.
The median votes per post was 5 and the average was 39.91. You can see how this compares to previous period below.
Flags
In votes above we also have flags, however combined only 0.99% of posts had a net downvote. This is down from 1.32% the previous month.
22K posts received 1 or more flags (this is posts not comments). This makes up 5.18% of all posts made. This is down from 9.34% the previous month.
In the first table we can see the authors with the highest % of flags and after this the table shows who gave the most flags. Following this we can see the time series for when flags were given
Other Data
The line chart above represents the number of new posts per day in September 18 and the bar chart below that shows the peak posting hours.
Conclusion
The number of posts each month has been on a downtrend this year and October was no exception. September seen the roll out of HF20 place restriction on posting and voting and in turn payouts, however we can see in October the payouts increase. So much so that the median payout on post for October was at an all time high for the year.
Engagement in terms of comments per post has remained rather consistent, however we seen a massive jump (over 100% jump) in the median number of votes per post.
The % of posts that received flags reduced by 53% on the previous period.
Keeping track of my own posts, on average I earned $29.30 per post, and each post received an average of 192 votes.
How did you posts do in October, please do leave a comment below and let me know how you got on
The Data and Queries
As always I have connected to Steemsql managed by @arcange using Power BI and the query is used was
SELECT
author,
category,
created,
children,
total_payout_value,
net_votes,
pending_payout_value
FROM
Comments (NOLOCK)
where created >= CONVERT(DATE,'2018-10-01')
and created< CONVERT(DATE,'2018-11-01')
and depth = 0
The code used to get flag details was
SELECT
*
FROM
txVotes (NOLOCK)
where timestamp >= CONVERT(DATE,'2018-10-01')
and timestamp< CONVERT(DATE,'2018-11-01')
and weight <0
Shameless Promotion
Like what we are doing? Dont forget to vote for @steemcommunity as your steem witness.