demarches-normaliennes/app/controllers/stats_controller.rb
Pierre de La Morinerie 503c393a87 helpscout: use replies_sent to compute contact rate
The `conversations_count` we previously used counts replies, but also
all conversations that were tagged or re-tagged during the month – with
made counting the actual work spent on user support brittle.

Counting the replies is a better estimation of what we get.

Unfortunately this also removes the filtering-by-tag feature. To
mitigate this, the reports are now scoped to a specific mailbox. This
allows to create extra mailboxes for conversations that should't be
counted in the stats.
2019-09-11 17:26:10 +02:00

447 lines
14 KiB
Ruby
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

class StatsController < ApplicationController
before_action :authenticate_administration!, only: [:download]
MEAN_NUMBER_OF_CHAMPS_IN_A_FORM = 24.0
def index
procedures = Procedure.publiees_ou_archivees
dossiers = Dossier.state_not_brouillon
@procedures_numbers = procedures_numbers(procedures)
@dossiers_numbers = dossiers_numbers(dossiers)
@satisfaction_usagers = satisfaction_usagers
@contact_percentage = contact_percentage
@dossiers_states = dossiers_states
@procedures_cumulative = cumulative_hash(procedures, :published_at)
@procedures_in_the_last_4_months = last_four_months_hash(procedures, :published_at)
@dossiers_cumulative = cumulative_hash(dossiers, :en_construction_at)
@dossiers_in_the_last_4_months = last_four_months_hash(dossiers, :en_construction_at)
if administration_signed_in?
@dossier_instruction_mean_time = Rails.cache.fetch("dossier_instruction_mean_time", expires_in: 1.day) do
dossier_instruction_mean_time(dossiers)
end
@dossier_filling_mean_time = Rails.cache.fetch("dossier_filling_mean_time", expires_in: 1.day) do
dossier_filling_mean_time(dossiers)
end
@avis_usage = avis_usage
@avis_average_answer_time = avis_average_answer_time
@avis_answer_percentages = avis_answer_percentages
@motivation_usage_dossier = motivation_usage_dossier
@motivation_usage_procedure = motivation_usage_procedure
@cloned_from_library_procedures_ratio = cloned_from_library_procedures_ratio
end
end
def download
headers = [
'ID du dossier',
'ID de la démarche',
'Nom de la démarche',
'ID utilisateur',
'Etat du fichier',
'Durée en brouillon',
'Durée en construction',
'Durée en instruction'
]
data = Dossier
.includes(:procedure, :user)
.in_batches
.flat_map do |dossiers|
dossiers
.pluck(
"dossiers.id",
"procedures.id",
"procedures.libelle",
"users.id",
"dossiers.state",
"dossiers.en_construction_at - dossiers.created_at",
"dossiers.en_instruction_at - dossiers.en_construction_at",
"dossiers.processed_at - dossiers.en_instruction_at"
)
end
respond_to do |format|
format.csv { send_data(SpreadsheetArchitect.to_xlsx(headers: headers, data: data), filename: "statistiques.csv") }
end
end
private
def procedures_numbers(procedures)
total = procedures.count
last_30_days_count = procedures.where(published_at: 1.month.ago..Time.zone.now).count
previous_count = procedures.where(published_at: 2.months.ago..1.month.ago).count
if previous_count != 0
evolution = (((last_30_days_count.to_f / previous_count) - 1) * 100).round(0)
else
evolution = 0
end
formatted_evolution = format("%+d", evolution)
{
total: total.to_s,
last_30_days_count: last_30_days_count.to_s,
evolution: formatted_evolution
}
end
def dossiers_numbers(dossiers)
total = dossiers.count
last_30_days_count = dossiers.where(en_construction_at: 1.month.ago..Time.zone.now).count
previous_count = dossiers.where(en_construction_at: 2.months.ago..1.month.ago).count
if previous_count != 0
evolution = (((last_30_days_count.to_f / previous_count) - 1) * 100).round(0)
else
evolution = 0
end
formatted_evolution = format("%+d", evolution)
{
total: total.to_s,
last_30_days_count: last_30_days_count.to_s,
evolution: formatted_evolution
}
end
def dossiers_states
{
'Brouilllon' => Dossier.state_brouillon.count,
'En construction' => Dossier.state_en_construction.count,
'En instruction' => Dossier.state_en_instruction.count,
'Terminé' => Dossier.state_termine.count
}
end
def satisfaction_usagers
legend = {
Feedback.ratings.fetch(:unhappy) => "Mécontents",
Feedback.ratings.fetch(:neutral) => "Neutres",
Feedback.ratings.fetch(:happy) => "Satisfaits"
}
number_of_weeks = 12
totals = Feedback
.group_by_week(:created_at, last: number_of_weeks, current: false)
.count
legend.keys.map do |rating|
data = Feedback
.where(rating: rating)
.group_by_week(:created_at, last: number_of_weeks, current: false)
.count
.map do |week, count|
total = totals[week]
# By default a week is displayed by the first day of the week but we'd rather display the last day
label = week.next_week
if total > 0
[label, (count.to_f / total * 100).round(2)]
else
[label, 0]
end
end.to_h
{
name: legend[rating],
data: data
}
end
end
def contact_percentage
number_of_months = 13
from = Date.today.prev_month(number_of_months)
to = Date.today.prev_month
adapter = Helpscout::UserConversationsAdapter.new(from, to)
if !adapter.can_fetch_reports?
return nil
end
adapter
.reports
.map do |monthly_report|
start_date = monthly_report[:start_date].to_time.localtime
end_date = monthly_report[:end_date].to_time.localtime
replies_count = monthly_report[:replies_sent]
dossiers_count = Dossier.where(en_construction_at: start_date..end_date).count
monthly_contact_percentage = replies_count.fdiv(dossiers_count || 1) * 100
[I18n.l(start_date, format: '%b %y'), monthly_contact_percentage.round(1)]
end
end
def cloned_from_library_procedures_ratio
[3.weeks.ago, 2.weeks.ago, 1.week.ago].map do |date|
min_date = date.beginning_of_week
max_date = min_date.end_of_week
all_procedures = Procedure.created_during(min_date..max_date)
cloned_from_library_procedures = all_procedures.cloned_from_library
denominator = [1, all_procedures.count].max
ratio = percentage(cloned_from_library_procedures.count, denominator)
[l(max_date, format: '%d/%m/%Y'), ratio]
end
end
def max_date
if administration_signed_in?
Time.zone.now
else
Time.zone.now.beginning_of_month - 1.second
end
end
def last_four_months_hash(association, date_attribute)
min_date = 3.months.ago.beginning_of_month.to_date
association
.where(date_attribute => min_date..max_date)
.group("DATE_TRUNC('month', #{date_attribute})")
.count
.to_a
.sort_by { |a| a[0] }
.map { |e| [I18n.l(e.first, format: "%B %Y"), e.last] }
end
def cumulative_hash(association, date_attribute)
sum = 0
association
.where("#{date_attribute} < ?", max_date)
.group("DATE_TRUNC('month', #{date_attribute})")
.count
.to_a
.sort_by { |a| a[0] }
.map { |x, y| { x => (sum += y) } }
.reduce({}, :merge)
end
def mean(collection)
(collection.sum.to_f / collection.size).round(2)
end
def percentage(numerator, denominator)
((numerator.to_f / denominator) * 100).round(2)
end
def dossier_instruction_mean_time(dossiers)
# In the 12 last months, we compute for each month
# the average time it took to instruct a dossier
# We compute monthly averages by first making an average per procedure
# and then computing the average for all the procedures
min_date = 11.months.ago
max_date = Time.zone.now.to_date
processed_dossiers = dossiers
.where(:processed_at => min_date..max_date)
.pluck(:groupe_instructeur_id, :en_construction_at, :processed_at)
# Group dossiers by month
processed_dossiers_by_month = processed_dossiers
.group_by do |dossier|
dossier[2].beginning_of_month.to_s
end
processed_dossiers_by_month.map do |month, value|
# Group the dossiers for this month by procedure
dossiers_grouped_by_groupe_instructeur = value.group_by { |dossier| dossier[0] }
# Compute the mean time for this procedure
procedure_processing_times = dossiers_grouped_by_groupe_instructeur.map do |_procedure_id, procedure_dossiers|
procedure_dossiers_processing_time = procedure_dossiers.map do |dossier|
(dossier[2] - dossier[1]).to_f / (3600 * 24)
end
mean(procedure_dossiers_processing_time)
end
# Compute the average mean time for all the procedures of this month
month_average = mean(procedure_processing_times)
[month, month_average]
end.to_h
end
def dossier_filling_mean_time(dossiers)
# In the 12 last months, we compute for each month
# the average time it took to fill a dossier
# We compute monthly averages by first making an average per procedure
# and then computing the average for all the procedures
# For each procedure, we normalize the data: the time is calculated
# for a 24 champs form (the current form mean length)
min_date = 11.months.ago
max_date = Time.zone.now.to_date
processed_dossiers = dossiers
.where(:processed_at => min_date..max_date)
.pluck(
:groupe_instructeur_id,
Arel.sql('EXTRACT(EPOCH FROM (en_construction_at - created_at)) / 60 AS processing_time'),
:processed_at
)
# Group dossiers by month
processed_dossiers_by_month = processed_dossiers
.group_by do |(*_, processed_at)|
processed_at.beginning_of_month.to_s
end
groupe_instructeur_ids = processed_dossiers.map { |gid, _, _| gid }.uniq
groupe_instructeurs = GroupeInstructeur.where(id: groupe_instructeur_ids).pluck(:id, :procedure_id)
procedure_id_type_de_champs_count = TypeDeChamp
.where(private: false)
.group(:procedure_id)
.count
groupe_instructeur_id_type_de_champs_count = groupe_instructeurs.reduce({}) do |acc, (gi_id, procedure_id)|
acc[gi_id] = procedure_id_type_de_champs_count[procedure_id]
acc
end
processed_dossiers_by_month.map do |month, dossier_plucks|
# Group the dossiers for this month by procedure
dossiers_grouped_by_groupe_instructeur = dossier_plucks.group_by { |(groupe_instructeur_id, *_)| groupe_instructeur_id }
# Compute the mean time for this procedure
procedure_processing_times = dossiers_grouped_by_groupe_instructeur.map do |groupe_instructeur_id, procedure_dossiers|
procedure_fields_count = groupe_instructeur_id_type_de_champs_count[groupe_instructeur_id]
if (procedure_fields_count == 0 || procedure_fields_count.nil?)
next
end
procedure_dossiers_processing_time = procedure_dossiers.map { |_, processing_time, _| processing_time }
procedure_mean = mean(procedure_dossiers_processing_time)
# We normalize the data for 24 fields
procedure_mean * (MEAN_NUMBER_OF_CHAMPS_IN_A_FORM / procedure_fields_count)
end
.compact
# Compute the average mean time for all the procedures of this month
month_average = mean(procedure_processing_times)
[month, month_average]
end.to_h
end
def avis_usage
[3.weeks.ago, 2.weeks.ago, 1.week.ago].map do |min_date|
max_date = min_date + 1.week
weekly_dossiers = Dossier.includes(:avis).where(created_at: min_date..max_date).to_a
weekly_dossiers_count = weekly_dossiers.count
if weekly_dossiers_count == 0
result = 0
else
weekly_dossier_with_avis_count = weekly_dossiers.select { |dossier| dossier.avis.present? }.count
result = percentage(weekly_dossier_with_avis_count, weekly_dossiers_count)
end
[min_date.to_i, result]
end
end
def avis_average_answer_time
[3.weeks.ago, 2.weeks.ago, 1.week.ago].map do |min_date|
max_date = min_date + 1.week
average = Avis.with_answer
.where(created_at: min_date..max_date)
.average("EXTRACT(EPOCH FROM avis.updated_at - avis.created_at) / 86400")
result = average ? average.to_f.round(2) : 0
[min_date.to_i, result]
end
end
def avis_answer_percentages
[3.weeks.ago, 2.weeks.ago, 1.week.ago].map do |min_date|
max_date = min_date + 1.week
weekly_avis = Avis.where(created_at: min_date..max_date)
weekly_avis_count = weekly_avis.count
if weekly_avis_count == 0
[min_date.to_i, 0]
else
answered_weekly_avis_count = weekly_avis.with_answer.count
result = percentage(answered_weekly_avis_count, weekly_avis_count)
[min_date.to_i, result]
end
end
end
def motivation_usage_dossier
[3.weeks.ago, 2.weeks.ago, 1.week.ago].map do |date|
min_date = date.beginning_of_week
max_date = date.end_of_week
weekly_termine_dossiers = Dossier.where(processed_at: min_date..max_date)
weekly_termine_dossiers_count = weekly_termine_dossiers.count
weekly_termine_dossiers_with_motivation_count = weekly_termine_dossiers.where.not(motivation: nil).count
if weekly_termine_dossiers_count == 0
result = 0
else
result = percentage(weekly_termine_dossiers_with_motivation_count, weekly_termine_dossiers_count)
end
[l(max_date, format: '%d/%m/%Y'), result]
end
end
def motivation_usage_procedure
[3.weeks.ago, 2.weeks.ago, 1.week.ago].map do |date|
min_date = date.beginning_of_week
max_date = date.end_of_week
procedures_with_dossier_processed_this_week = Procedure
.joins(:dossiers)
.where(dossiers: { processed_at: min_date..max_date })
procedures_with_dossier_processed_this_week_count = procedures_with_dossier_processed_this_week
.uniq
.count
procedures_with_dossier_processed_this_week_and_with_motivation_count = procedures_with_dossier_processed_this_week
.where
.not(dossiers: { motivation: nil })
.uniq
.count
if procedures_with_dossier_processed_this_week_count == 0
result = 0
else
result = percentage(procedures_with_dossier_processed_this_week_and_with_motivation_count, procedures_with_dossier_processed_this_week_count)
end
[l(max_date, format: '%d/%m/%Y'), result]
end
end
end