Which witnesses could be interesting for me?

Voting for witnesses is very important. Especially after the HF20 failure of the top 20 witnesses, everyone should take a look over their witness votes another time. But the witness list is long, and checking each witness in depth would take a huge time. That's why I wrote a script to check for which witnesses the persons vote for, that you follow. That gives you a list of witnesses that it is definitely worth checking. This isn't a list of witnesses that you should vote for, but a list that you should look into. Also, there are two values for each witness in the list. One is the number of followings that vote for this witness (the list is sorted by that) and the other is a special score. For each followed account, 30/number of witness votes from that account gets added to the score. That causes an account that only votes for @reggaemuffin (a cool witness by the way) to add a lot more to @reggaemuffin's score than an account that votes for @reggaemuffin and 29 other witnesses. The witnesses you are already voting for, are excluded for the list. This approach is very similar to the approach @holger80 (another awesome witness) had for looking for interesting persons to follow. Look at his post if you are interested in that. The code looks like this:

from beem.account import Account
from beem.nodelist import NodeList
from beem.steem import Steem
from beem.witness import Witness

if __name__ == '__main__':
    nodes = NodeList()
    nodes.update_nodes()
    node_list = nodes.get_nodes()
    stm = Steem(node=node_list)
    main_acc = Account('flugschwein', steem_instance=stm)

    following = main_acc.get_following()
    main_witness_votes = stm.rpc.get_account(main_acc.identifier)[0]['witness_votes']
    witnesses_weighted = {}
    witnesses_straight = {}
    for a in list(set(following + main_witness_votes)):
        acc = stm.rpc.get_account(a)[0]
        for w in acc['witness_votes']:
            if w in main_witness_votes:
                continue
            elif w in witnesses_weighted:
                witnesses_weighted[w] += 30 / acc['witnesses_voted_for']
                witnesses_straight[w] += 1
            else:
                witnesses_weighted[w] = 30 / acc['witnesses_voted_for']
                witnesses_straight[w] = 1
    final_witnesses = []
    for w in witnesses_weighted:
        wit = Witness(w)
        final_witnesses.append(
            {'name': wit.identifier, 'score': witnesses_weighted[w], 'url': wit['url'], 'cnt': witnesses_straight[w]})
    final_witnesses = sorted(final_witnesses, key=lambda x: x['cnt'], reverse=True)
    print('|My Witness Votes|My following\'s Witness Votes|')
    print('|---|---|')
    print(f'|{len(main_witness_votes)}|{len(final_witnesses)}|')
    print('\n|Number|Name|Weighted Score|Count|')
    print('|---|---|---|---|')
    for i in range(30):
        print('|{}|[{}]({})|{}|{}|'.format(i + 1, final_witnesses[i]['name'], final_witnesses[i]['url'],
                                           round(final_witnesses[i]['score'], 3), final_witnesses[i]['cnt']))

This script still has quite some room for improvement, but that would only affect performance. I will keep the script in here up to date. For me, @flugschwein, that script returns the following: |My Witness Votes|My following's Witness Votes| |---|---| |18|129|