php 43 lines · 8 steps

Computing rolling averages in Laravel

A service class turns daily metric rows into a per-day series with a trailing N-day rolling average.

Explained by highlit

Laravel Intermediate rolling-average date-ranges collections time-series eloquent

1<?php
2 
3namespace App\Services;
4 
5use App\Models\DailyMetric;
6use Carbon\CarbonPeriod;
7use Illuminate\Support\Collection;
8 
9class RollingAverageCalculator
10{
11    public function __construct(private int $window = 7)
12    {
13    }
14 
15    public function forRange(string $metric, \DateTimeInterface $from, \DateTimeInterface $to): Collection
16    {
17        $totals = DailyMetric::query()
18            ->where('name', $metric)
19            ->whereBetween('recorded_on', [$from, $to])
20            ->pluck('value', 'recorded_on')
21            ->mapWithKeys(fn ($value, $date) => [substr($date, 0, 10) => (float) $value]);
22 
23        return collect(CarbonPeriod::create($from, $to))
24            ->map(function ($day) use ($totals) {
25                $date = $day->toDateString();
26                $windowStart = $day->copy()->subDays($this->window - 1);
27 
28                $values = collect(CarbonPeriod::create($windowStart, $day))
29                    ->map(fn ($d) => $totals->get($d->toDateString()))
30                    ->filter(fn ($v) => $v !== null);
31 
32                return [
33                    'date' => $date,
34                    'value' => $totals->get($date, 0.0),
35                    'rolling_average' => $values->isEmpty()
36                        ? null
37                        : round($values->avg(), 2),
38                    'samples' => $values->count(),
39                ];
40            })
41            ->values();
42    }
43}

01 / 01

STEP 01

Walkthrough

Space play ←→ step click any line

Three takeaways

1Loading raw values into a date-keyed map once lets you compute every window with fast in-memory lookups instead of repeated queries.
2Iterating a CarbonPeriod gives you a gap-free day-by-day spine even when the underlying data has missing days.
3Tracking the sample count alongside the average makes partial or sparse windows honest rather than silently misleading.

Related explainers

php

<?php
 
namespace App\Console\Commands;

Releasing stale document locks in Laravel

artisan-command transactions row-locking

Intermediate 6 steps

php

<?php
 
namespace App\Providers;

Subdomain multi-tenancy routing in Laravel

multi-tenancy service-container route-binding

Advanced 7 steps

php

<?php
 
namespace App\Broadcasting;

Authorizing presence channels in Laravel

broadcasting authorization presence-channels

Intermediate 3 steps

java

@RestController
@RequestMapping("/api/products")
public class ProductSearchController {

Binding collection query params in Spring

rest-api query-parameters dependency-injection

Intermediate 6 steps

php

<?php
 
namespace App\Http\Middleware;

Resolving the current team in Laravel middleware

middleware multi-tenancy cookies

Intermediate 8 steps

php

<?php
 
namespace App\Http\Controllers;

Broadcasting typing indicators in Laravel

websockets authorization presence-channels

Intermediate 9 steps

Share this explainer

Here's the card — post it anywhere.

Computing rolling averages in Laravel — share card

Made with highlit — turn any snippet into a walkthrough like this in about a minute.

Explain your code