What if Your Post Spread Like a Virus!? Looking for a SteemDev.

This is the very thing that I am trying to accomplish. The idea is simple. When someone upvotes your post, it will show up in the feed of people that vote similarly to whoever upvoted it. In this way, it will spread through the feeds of people who have similar interests.

I explain my initial idea in this post, and put out an initial budget proposal in this post if you want more details.

I wanted to give a quick update.

I had to take a break for a few weeks because there was an update that to the way I was accessing the steem blockchain that broke my method which made it impossible to do any more research, but luckily it is working again so I have been testing a few different algorithms.

The method I am testing now will give a score to each account that you vote similarly to. The higher the score, the more impact their future votes will have on your feed. This is how I get the initial scores:

1. Get my last 100 votes (I would like to do my votes for the last 7 days but that is above my ability at this point)
2. Get everyone who has voted on those 100 posts
3. Count how many times each account voted on the same posts I did
4. Take the top 40 accounts (That is about all my spreadsheet can handle) and get at the last 100 posts each of them voted on.
5. Create my feed by getting the posts that appeared the most in step 4.
6. Browse through the feed and upvote any posts that you like and skip the rest.
7. Repeat Step 1
8. Get the account score for each account:
   8a. Count how many posts you both voted on in the last 100 votes (voted)
   8b. Count how many posts you skipped that they voted on (skipped)
   8c. Calculate the score with the formula as follows => voted/(voted+skipped)
   example:
   While I browsed through my feed, I upvoted 4 posts that John had upvoted and skipped 6 that he had voted on. Taking the formula above 4/(4+6) = .4
   John's score is .4

After we get the initial scores we adjust our method slightly. Instead of using the top 40 people that voted similar like last time we'll take the top 5 of them (excluding any that already have scores) as well as the 35 accounts with the highest scores. These are the steps

1. Get my last 100 votes

2. Get everyone who has voted on those 100 posts

3. Count how many times each account voted on the same posts I did

4. Take the top 5 accounts from this list (excluding any accounts that already have scores) and the 35 accounts with the highest scores and get at the last 100 posts each of them voted on. (We will call these accounts "chosen" for this iteration)

5. Create my feed by giving each post a score and putting them in order from highest score to lowest
   5a. The "post score" is calculated by taking the sum of the score of each "chosen" account that voted on it. Example:
   Our example post was voted on by John, Matt, and Pam. Their scores were .4, .15, and .25 respectively. Our post score would be .4+.15+.25 which equals .8

6. Browse through the feed and upvote any posts that you like and skip the rest.

7. Repeat Step 1

8. Get the account score for each account (notice that the formula is slightly different because we add the new score to the old scores and divide it be the times chosen):
   8a. Count how many posts you both voted on in the last 100 votes (voted)
   8b. Count how many posts you skipped that they voted on (skipped)
   8c. Calculate the score with the formula as follows => (voted/(voted+skipped)+ all old scores)/times chosen (this gives us an average score for each time the account is chosen)

At this point we just repeat these steps indefinitely, over time it will start to identify the accounts that vote the most similar to you and prioritize posts they vote on in your feed.

I have been testing this method for several days now and I have enjoyed finding posts that I might never have found otherwise. I'll keep you up to date on any changes I make and how it works out.

If anyone is interested in helping me build this into a webapp or possibly an extension of some kind let me know. You can reply to this post or contact me on steemit.chat. Once I can price out how much it would cost to build I can re-submit it as a budget proposal to fund the development.

Any suggestions are appreciated.