Ruby on Rails Demographic Data

Question

I made an site for a PS3 game and I have quite a lot of users. I am wanting to make tournaments based on peoples locations and would also like to target age groups. When users sign up the input there date of birth in the format YYYY-MM-DD. I am pulling the data and making it into a hash like so:

# Site.rb

has_many :members

def ages
  ages = {"Under 18" => 0, "19-24" => 0, "25-35" => 0, "36-50" => 0, "51-69" => 0,"70+" => 0}
  ages_results = self.members.count("DATE_FORMAT(dob, '%Y')", :group =>"DATE_FORMAT(dob, '%Y')")
  ages_results.each do |k,v|
    k = k.to_i
    if k.between?(18.years.ago.strftime("%Y").to_i, 0.years.ago.strftime("%Y").to_i)
      ages["Under 18"] += v
    elsif k.between?(24.years.ago.strftime("%Y").to_i, 19.years.ago.strftime("%Y").to_i)
      ages["19-24"] += v
    elsif k.between?(35.years.ago.strftime("%Y").to_i, 25.years.ago.strftime("%Y").to_i)
      ages["25-35"] += v
    elsif k.between?(50.years.ago.strftime("%Y").to_i, 36.years.ago.strftime("%Y").to_i)
      ages["36-50"] += v
    elsif k.between?(69.years.ago.strftime("%Y").to_i, 51.years.ago.strftime("%Y").to_i)
      ages["51-69"] += v 
    elsif k > 70.years.ago.strftime("%Y").to_i
      ages["70+"] += v
    end
  end
  ages
end

I am not a expert ruby developer and not sure if the above approach is good or it can be done a much better way, could anyone give me some advice about this?

Cheers

score 3 · Accepted Answer · edited May 23 '17 at 11:44

Couple of things to note in your code:

you seem to disregard month and day when a user was born
you convert to and from strings unnecessarilly:
```
50.years.ago.strftime("%Y").to_i
```
could be written as
```
50.years.ago.year
```
hard-coded values all over the code

I would start rewriting by finding an adequate method for calculating exact age. This one seems to be ok:

require 'date'

def age(dob)
  now = Time.now.utc.to_date
  now.year - dob.year - ((now.month > dob.month || (now.month == dob.month && now.day >= dob.day)) ? 0 : 1)
end

Then I would extract age table to a separate structure, to be able to change it easily, if needed, and have it visually together:

INF = 1/0.0 # convenient infinity
age_groups = {
  (0..18) => 'Under 18',
  (19..24) => '19-24',
  (25..35) => '25-35',
  (36..50) => '36-50',
  (51..69) => '51-69',
  (70..INF) => '70+'
}

Next you can take as the input the array of users' birth dates:

users_dobs = [Date.new(1978,4,16), Date.new(2001,6,13), Date.new(1980,10,22)]

And starting to find a suitable method to group them based on your map, say using inject:

p users_dobs.each_with_object({}) {|dob, result|
  age_group = age_groups.keys.find{|ag| ag === age(dob)}
  result[age_group] ||= 0
  result[age_group] += 1
}
#=>{25..35=>2, 0..18=>1}

or, perhaps, using group_by

p users_dobs.group_by{|dob|
  age_groups.keys.find{|ag| ag === age(dob)}
}.map{|k,v| [age_groups[k], v.count]}
#=>[["25-35", 2], ["Under 18", 1]]

etc.

Thanks for putting the effort into the answer, it has helped make my code more elegant! — RailsSon, Sep 19 '11 at 13:05

Ruby on Rails Demographic Data

1 Answers1