Company twitter and the Pareto principle

From wikipedia: “The Pareto principle (also known as the 80-20 rule,[1] the law of the vital few, and the principle of factor sparsity) states that, for many events, roughly 80% of the effects come from 20% of the causes.”

I’ve always liked the 80-20 rule, because it seems to be such a consistent rule of thumb when dealing with people, who you would assume would be more random.

So recently we had our company update, and a chart was put up showing who in the company are the biggest tweeters, which made me wonder how closely that data would follow the 80-20 rule, given how small the sample size is. So here we go.

First the code. I think it would have taken me longer to find someone elses program that does this than it took me to write it:

First the code:

require 'rubygems'
require 'twitter'
require 'terminal-table/import'

def percentage(tweets)
  "%.2f" % (tweets.to_f/@total_tweets*100) + '%'
end

httpauth = Twitter::HTTPAuth.new('C3Beau', 'xxxxxxxx')
client = Twitter::Base.new(httpauth)

c3users = client.followers << client.user('C3Beau')

tweet_counts = c3users.collect {|user| user[:statuses_count]}
@total_tweets = tweet_counts.inject(0) {|sum,count| sum + count }

c3users.sort! {|a,b| a[:statuses_count] <=> b[:statuses_count] }
output_table = c3users.collect do |user|
  [
    user[:screen_name],
    user[:statuses_count],
    percentage(user[:statuses_count])
  ]
end

puts table(['user','tweets','percentage'],*output_table)

Next the output:

+-----------+--------+------------+
| user      | tweets | percentage |
+-----------+--------+------------+
| C3Burgo   | 0      | 0.00%      |
| C3Krishni | 0      | 0.00%      |
| C3Sreyna  | 1      | 0.07%      |
| c3deepika | 1      | 0.07%      |
| C3Seanc   | 1      | 0.07%      |
| C3Blair   | 4      | 0.29%      |
| C3Zelman  | 4      | 0.29%      |
| C3Kathy   | 5      | 0.36%      |
| C3Tony    | 9      | 0.65%      |
| C3Rob     | 11     | 0.79%      |
| C3SimonC  | 13     | 0.94%      |
| C3Tara    | 22     | 1.59%      |
| C3Tim     | 28     | 2.02%      |
| C3Sumo    | 38     | 2.74%      |
| C3ChrisM  | 74     | 5.33%      |
| C3Jodie   | 83     | 5.98%      |
| C3Tak     | 93     | 6.70%      |
| C3Mark    | 104    | 7.49%      |
| C3Michael | 105    | 7.56%      |
| C3Cameron | 159    | 11.46%     |
| C3Beau    | 169    | 12.18%     |
| C3Conrad  | 204    | 14.70%     |
| C3Alister | 260    | 18.73%     |
+-----------+--------+------------+

So, working from highest to lowest until we get to around 80% we go.. 18.73+14.70+12.18+11.46+7.56+7.49+6.70 = 78.82%

That got us from C3Alister to C3Tak. 7 People, out of a sample size of 23 giving us… 30%. Ha, considering there’s only 23 people in the group I reckon that’s pretty good.

Note: All of those accounts are protected, so I figure I’m safe in listing them like that.

Notes