Steemit.com - The Impact of new accounts on daily number of posts

Introduction

It is essential for any kind of social media platform to increase the number of interactions on the platform besides attracting new users who are the real agents behind those valuable interactions. The conventional social media platforms instrumentalize those two fundamental properties to receive more benefits from the advertisements. Steem Blockchain's revolutionary system design fundamentally changed the rules of the game. With the fascinating discovery of PoB (Proof-of-Brain) mining system, steem blockchain has the capacity to facilitate any sort of social media interaction which a standard user does on the internet as valid method of cryptocurrency mining.
With regard to Steem's way of handling the mining issue, it is a necessity to check the relationship between the very activity of mining and the number of new miners in the field. This article aims to clarify the impacts of new miners (new accounts) on daily mining (number of posts per day -as one indicator).

Theoretical Background & Hypothesis

Accordint to self-interst theory, minimization of the costs and maximization of the benefits are the two main motivations behind the every human action. It's a fact that every created account starts with 0 followers and in most of the cases users do not reach significant numbers of followers in short notice. In such cases of low numbers of followers, it is not cost-efficient for a new user to produce high quality contents which increases the opportunity cost while bringing the same levels of benefits. The best way to become cost efficient for the users is to decrease the time they spend on production of the content and to increase the number of the posts at the same time. When we evaluate this fact in line with the explanatory framework of self-interest theory, we may form the following hypothesis about steemit.com's user behavior:

• The daily number of the new acounts positively effects the daily number of posts.

It can be considered as a safe hypothesis but testing this hypothesis by simple linear regression will provide us a crucial information about the strenght of this impact in numbers.

Scope of Analysis & Limitations

The scope of the analysis is 12 months time period between 2017/01/19-2018/01/19.
The article is designed to provide ceteris paribus interaction between two variables.
Since the demographic variables (age, gender, education) are not available, this analysis has no control variables.

Data, Methods & Operationalization

In order to get the data from the steem blockchain i used sql querries.
The Sql server
I used the help of this magnificent steemhelper:
Sql.steemhelper
I used Valentina Studio to run the querries and get the data in excel format.
To get the number created accounts per each day in the last year:

SELECT CONVERT(DATE, timestamp) as Date, COUNT(*) as [New Accounts]
FROM TxAccountCreates
WHERE timestamp >= '2017/01/19' AND timestamp < '2018/01/19'
GROUP BY CONVERT(DATE, timestamp)
ORDER BY CONVERT(DATE, timestamp) DESC

To get the number of posts per each day in the last year:

SELECT CONVERT(DATE, created) as Date, COUNT(*) as Posts
FROM Comments
WHERE created >= '2017/01/19' AND timestamp < '2018/01/19'
GROUP BY CONVERT(DATE, created)
ORDER BY CONVERT(DATE, created) DESC

After achieving the the data, i used Stata 13.0 to run the Simple Linear Regression
Steps to run the linear regression

1. Merging the datasets
2. Renaming the variables as posts(dependent variable) accounts (independent variable) {optional}
3. Running the SLR by:

reg posts accounts

a) Getting the scatter plot visualization by

twoway (scatter posts accounts, sort)

b) Getting the prediction line

predict PredictedPost, xb
twoway (scatter posts accounts, sort) (line PredictedPost accounts)

Results & Interpretation

Regression:

In order to check the significance level, firts we have to check the P>|t| value:
The value is 0.000 and it means the results are extremely significant. As a next step we should find the coefficient value on the table to understand the direction and the stenght of the relation. There we see "25.27006". This number means every one unit increase in the number of created accounts, will entail a 25.27006 increase in the number of posts per day. Standard error gives us the possible fluctuations in this number. Our standard error is "1.379885" which means every increase in the number of the created accounts per day brings an increase between 23.890175 and 26.649945. Roughly speaking it is between 24 and 26.
By relying on this we can say our hypothesis is proven correct.

Scatter plot:

Here the each plot is representing a day. As we can see , the plots are concentraded as groups. There are also few outlier cases which could be explained by multivariate analysis. For practical reasons explaining them is out of the scope of this paper. The general trend of the dots is visible for many of us. We can draw a line to point that trend. But there is a more scientific way to do this. Stata draws it for us.

This lines is called linear prediction line. It's main function is to give us linear line which cuts the plots into two groups. This line is called prediction line because it also provides us a future projection about numbers. 8638 new accounts per day will entail posting of 250000 posts in that day.

Repharasing all with the technical terminology, this study show us the "new miner/mining measured by posts" ratio of the steemit.

Best Regards,

---

Posted on Utopian.io - Rewarding Open Source Contributors

---