This was a solution to a problem I was trying to solve on my webiste: how can I arrive at an average value over time, when new data is being provided, without having to store every datum.
I wrote this code to prove to myself that I wasn't crazy, and that this was indeed a valid way to calculate the mean.
| # Collect 100 random data | |
| values = 100.times.map { rand(100) } | |
| # This is the original average of the values | |
| start_average = values.inject(:+) / 100.0 | |
| # Print out to the console to prove you're not crazy | |
| puts "Start average:" | |
| puts start_average | |
| # Add another value to the array to get a new average value. This simulates the case where every data point is stored, and | |
| # the average is calculated the "old fashioned way", with every data available. | |
| puts "New average with extra datum:" | |
| puts (values + [200]).inject(:+) / 101.0 | |
| # Can't stop creating objects for everything :neckbeard: | |
| class NewAverageCalculator | |
| attr_accessor :new_value, :data_count, :current_average | |
| # Create a new calculator. | |
| # | |
| # new_value - The new datum added to the set | |
| # current_average - The previously calculated average value of all data | |
| # data_count - The total number of data previously collected | |
| def initialize(new_value, current_average, data_count) | |
| @average = current_average | |
| @index = data_count | |
| agreggate_average(new_value) | |
| end | |
| # The new average value | |
| # | |
| # Returns a Float | |
| def new_average | |
| @new_average | |
| end | |
| private | |
| def agreggate_average(new_value) | |
| @new_average = Rational((@average * @index) + new_value, @index + 1).to_f | |
| @index += 1 | |
| return @new_average | |
| end | |
| end | |
| puts "Current average:" | |
| puts NewAverageCalculator.new(200, start_average, 100).new_average |