DApps on STEEM - Analyzing SteemHunt

  Steemhunt Discover Cool Products and Get Rewards by STEEM. Product hunt on STEEM

Repository 

https://github.com/Steemhunt

@steemhunt has been in receipt of funding from Steemit Inc by way of delegation from @Misterdelegation.  As one of the newer apps on the block, this is the first analysis I have published on data around @steemhunt, and I think it is a project worth showcasing.

In addition to delegation from @misterdelegation, steemhunt has also raised delegation of circa ~600KSP from other sources.

Aim of Analysis

• Obtain a general feel for the activity levels on steemhunt

• Find out how much was paid out in rewards by steemhunt 

• Find out how much posts to steemhunt earned in total

• See who the movers and shakers are

• Correlation tests on voting habits

• User profiling

Scope of Analysis

The analysis was carried out for the month of October 18.  Data from the STEEM blockchain and Coinmarketcap was used and full details of the queries can be found at the bottom of this post.

The value of a vote for steemhunt was taken from steemworld.org on 10 Nov.

Findings

• 57% of steemhunt posts and comments received a vote from @steemhunt. 

• @steemhunt votes make up 83% of the gross payouts on steemhunt posts.

• Over US$30.59 in gross payouts was earning by authors using steemhunt.

• positive correlation (0.5546) was found between the % weight of a vote received and the owned SP by the author.

• The user based voted on has relatively the same profiling as the user base posting with the exception of % of posts made via steemhunt relative to all posts made,

Activity on Steemhunt

  

18.5K posts were made by 980 authors and received 55K  comments from 1826 authors and the comments grossed a payout of SBD$11.14  from 86K votes.

The posts also netted over 399K votes with a gross payout of SBD$19.67K.

In US$ this equates to $30.59  in gross payouts.

 

Votes From Steemhunt

  

724 distinct Authors on both posts and comments, received in total 36K votes from Steemhunt in October.  The average vote weight was 0.96% giving a vote of approx. 1 SBD.  In total Steemhunt have funded 25.56K SBD by way of upvotes.

If we look at this in US$ we can see that steemhunt votes have awarded payouts to the tune of US$25.30K

  

57% of steemhunt posts and comments received a vote from @steemhunt. 

@steemhunt votes make up 83% of the funding earning on steemhunt posts and comments

Movers and Shakers

The table below shows the top hunters by way of steemhunt votes received

  

The most active posters are shown in the table below

  

The table below shows the hunters most active in comments

  

Correlation tests on Steemhunt Voting

A series of tests were carried out to see if there is a correlation between steemhunt voting 

The first test is to see if there is a correlation between the Average voting weight used by steemhunt and the % of app posts submitted per author that received a vote.  A correlation coefficient of -0.107 was found.  Therefore we can establish there is no / very weak negative correlation between the voting weight and the % of your posts that get approved.

  

The second test is to see if there is a correlation between the voting weight used and the % of total posts by an author on STEEM that relate to steemhunt.  A correlation coefficient of -0.172 was found which indicates no correlation

The third test is to see if there is a correlation between the owned SP of the posting account and the average weight used by steemhunt on the voting.  It appears that there is a positive correlation (0.5546) between these values, which is heavily influenced by 3 outliers. Removing these outliers, we would get a correlation of 0.2.

  

The forth test looks for a correlation between the voting weight used and the age of the account that is posting the project.  An extremely weak correlation can be found.

User Profiling

I thought it would be interesting to take a look at the user profiling of both those supported by Steemhunt  and those posting to steemhunt

The Green represents accounts that have received a vote from steemhunt and the blue include all authors posting to steemhunt.

The below histogram shows the % of posts an account made to steemhunt against all their posts on steem.  297 account made >96% of their posts to steemhunt in October.  181 of these were supported.  On the other end of the scale 865 people had <=16% of their steem posts to steemhunt and 86 of these were supported.  From this we can tell that steemhunt tend to favour loyal users.

  

The below histogram shows the age of the STEEM account posting to steemhunt. The green showing the spread of account voted for and the blue showing the spread of the age of accounts that posted to steemhunt.  Both histograms follow the same pattern 

  

Above we can see the accounts by SP owned and the distribution of both posts and votes seems to be very much in line with each other and the vast majority of accounts have active have 29KSP or less

Below shows the distribution of steemhunt users in terms of powering down

  

Above shows the distribution of users in terms of those that power up liquid steem.

Conclusion

Steemhunt is an amazing project allowing anyone search fro awesome products and share them on the block.  I would very much encourage you to take a look at the project.

There is a positive correlation between the SP owned and the Average weight used when voting by steemhunt.  The outliers in this case are as follows

 

It is also interesting to see with the user profiling, the more exclusive your posts on steem are to steemhunt the more likely it will receive a vote.  This suggests they are rewarding loyalty.

During this analysis I also found that users rewarded with have powered down 49k SP while on the other side, users have power up 39K liquid steem in addition to SP growth from Authors and Curators rewards. 

Shameless Promotion

 Like what we are doing?  Dont forget to vote for @steemcommunity as your steem witness. 

Data and queries

Data was gathered in Power BI and Excel From Steemsql held and managed by @arcange.

The M code used to get vote information from steemhunt was

let

    Source = Sql.Database("vip.steemsql.com", "DBSteem", [Query="SELECT#(lf)*#(lf)FROM #(lf)Txvotes (NOLOCK)#(lf)#(lf)WHERE [voter] in ('steemhunt')#(lf)and CONVERT(DATE,timestamp) BETWEEN '2018-10-01' AND '2018-10-31'"]),

    #"Added Custom" = Table.AddColumn(Source, "% weight", each [weight]/10000),

    #"Changed Type" = Table.TransformColumnTypes(#"Added Custom",{{"% weight", Percentage.Type}}),

    #"Added Custom1" = Table.AddColumn(#"Changed Type", "Current vote worth", each 95.44),

    #"Added Custom2" = Table.AddColumn(#"Added Custom1", "approx vote $", each [Current vote worth]*[#"% weight"]),

    #"Changed Type1" = Table.TransformColumnTypes(#"Added Custom2",{{"timestamp", type date}, {"approx vote $", type number}})

in

    #"Changed Type1"

The M code used to post information to steemhunt was

let

    Source = Sql.Database("vip.steemsql.com", "DBSteem", [Query="SELECT#(lf)    author,#(lf)json_metadata,#(lf) permlink,#(lf)      total_payout_value,#(lf)    pending_payout_value,#(lf)    net_votes,#(lf)    created,#(lf)    children,#(lf)    Category,#(lf)    Depth#(lf)#(lf)FROM comments (NOLOCK)#(lf)where CONVERT(DATE,created) BETWEEN '2018-10-01' AND '2018-10-31' and#(lf)contains (json_metadata, 'steemhunt')#(lf)#(lf)#(lf)#(lf)"]),

    #"Changed Type" = Table.TransformColumnTypes(Source,{{"created", type date}})

in

    #"Changed Type"

The M code used to get coinmarketcap data was

let

    Source = Web.Page(Web.Contents("https://coinmarketcap.com/currencies/steem-dollars/historical-data/?start=20180101&end=20181031")),

    Data0 = Source{0}[Data],

    #"Changed Type" = Table.TransformColumnTypes(Data0,{{"Date", type date}, {"Open*", type number}, {"High", type number}, {"Low", type number}, {"Close**", type number}, {"Volume", type number}, {"Market Cap", type number}}),

    #"Removed Other Columns" = Table.SelectColumns(#"Changed Type",{"Date", "Close**"})

in

    #"Removed Other Columns"