About SNS account value calculation algorithm

I'd like to discuss algorithms, but it's a little abstract. I'm not talking about pure programming technology. I don't know if it's a good question to ask here.
(If it doesn't work, I'll turn it off.)

I would like to quantify the account quality of SNS

Take Twitter as an example.

Follows Follows Follows Start Date and Time

3 of
is readily available Tweets, replies, RTs, and so on. You can also calculate the number of Favs and RTs for individual tweets

From this information
for that account Value=Value for outgoing tweets, benefits to follow
I'd like to calculate that

*Account quality here is the value of the information to be transmitted
"""Notes and influences = How many people will see it if you let that person advertise""

It's a little different from

Even if you have just registered or have few followers, I'd like to calculate things like secretly sending high-value information

I think we'll be able to match over time, but
I want to judge the value of an account that is growing as soon as possible

judging purely by the number of followers
For example, tweets have the same value, but
For those who only follow back and those who follow themselves,
The number of followers is different, isn't it?

If you don't do anything, you can follow the same number of people at the same time. How many people will unconditionally return me as someone I follow?
Tweet quality is the same, but the number of followers is different

However, Even if you use the difference or ratio of the number of followers,
There is a difference in the number of followers even if the number of followers is the same, not the person who follows you back. If you use the difference or ratio, the value of the person who follows you back will decrease

For Fav and RT, it basically depends on the number of followers, so
Tweets of the same value will increase depending on the number of followers
(As for Twitter, it will spread over time and become a value standard to some extent.)

Also, if you come up with individual tweets, you can only occasionally tweet with pure value
Sometimes I say good things, but there are also people who throw away a lot of low-value tweets.

Some people say good things sometimes, and others say that it's okay to tweet anything.
Some people think worthless tweets are a hindrance.
Since it is difficult to calculate it uniformly,
"As for ""every time I want to follow"", I would like to think about it based on the number of followers, but
If you use only the number of followers
As I mentioned earlier, there is a problem that there is a difference in frequency of follow-back and active follow-up.

The best part is
Number of previously followed/Account Duration
I wish I knew something like
Time information for follow-up followers cannot be retrieved

Are there any calculation indicators or papers that are already in use?
Or if you have any ideas, please let me know

I haven't decided on a specific purpose yet at the planning stage. For now, I'm thinking vaguely about hiring personnel and using appropriate business decisions.
The calculation time seems to be okay to take a few hours to a day per account.
(I haven't investigated how much I can load the Twitter server to trace each tweet back, but I would like to calculate it with information that I can get without bothering you.)

Applicationally, judging spam stains, etc.
I think it can also be used to introduce recommended accounts.
In that case, I would be even happier if I could calculate for a large number of accounts in a little less time