How to aggregate barchart data for volumes of registrations per day from Django auth?

Question

Django auth User model has a date_joined DateTimeField. Could this be used to aggregate a list with volumes of registrations per day for days (or other time periods) in a daterange? E.g.: [(21.01, 5), (22.01, 7), (23.01, 9), ...] What is the fastest way to do this? E.g. if the date range on the barchart was set to past 3 years.

I looked at http://docs.djangoproject.com/en/dev/topics/db/aggregation/, but it's not apparent how to breakdown objects using their timestamps. This is a common design pattern in statistics.

score 1 · Answer 1 · edited May 23 '17 at 11:55

1

This question is similar to this one: django: time range based aggregate query

Something like this should work: MyObject.objects.filter(date_joined__date__range=(weekago,today)).annotate(registrations=Count('id')

edited May 23 '17 at 11:55

Community

1
1

answered Jan 21 '11 at 04:38

Lyle Pratt

5,636
4
27
28

Not really, I'll get the datetime `from datetime import datetime, timedelta` and User model `from django.contrib.auth.models import User`, now I select past 90 days `past90d = (datetime.now() - timedelta(days=30), datetime.now())` **now I'll try your tip**: `signups = User.objects.filter(date_joined__date__range = past90d).annotate(registrations = Count('id'))`. This raises: **join not permitted on date_joined field**. Ok, I'll take out `date__range=` leaving just `__range=`. No dice, this returns a list of user objects instead of expected list of days with registration counts. Any ideas? – fmalina Jan 31 '11 at 17:22

score 0 · Accepted Answer · answered Feb 02 '11 at 19:45

Here's my code for generating date-based barcharts with Django and CSS.

def date_barchart(model, date_field, operation='count', days=150, till=datetime.now()):   
    list = []
    i = 0
    while i < days:
        day = (till - timedelta(days=i), till - timedelta(days=i-1))
        date = time.strftime("%Y-%m-%d", day[0].timetuple())
        kwargs = {'%s__range' % date_field: day}
        qs = model.objects.filter(**kwargs)
        if operation == 'count':
            total = qs.count()
        else:
            total = qs.aggregate(Sum('amount'))
            total = total['amount__sum']
        if total > 0:
            list.insert(0, {'date': date, 'total': total, 'round_total': int(total)})

        i = i + 1
    return list

In my view I add may barchart data to template's context:

'barchart': date_barchart(User, 'date_joined')

or for Satchmo e-commerce

'barchart': date_barchart(Payment, 'time_stamp', 'sum', 150)

template:

{% if barchart %}
<div class="barchart">
    <ol>
        {% for d in barchart %}
        <li><u style="height:{{ d.round_total }}px"><i>{{ d.date }}<br>
            <b>{{ d.total }}</b></i></u></li>
        {% endfor %}
    </ol>
</div>
{% endif %}

and CSS:

.barchart{width:100%;overflow:hidden}
.barchart u{display:inline-block;position:relative;vertical-align:bottom;width:5px;background:#6cc;border-right:1px solid #fff;text-decoration:none}
.barchart u i{display:none;position:absolute;background:#fff;border:1px solid #aaa;padding:5px;border-radius:5px;left:-6px;bottom:10px;width:60px;font-size:10px;font-style:normal;color:#000}
.barchart u:hover{background-color:#333}
.barchart u:hover i{display:inline-block;z-index:500}
.barchart li{margin:0;padding:0;list-style:none;display:inline-block}

Looks as good as Google analytics, but no need to give away my performance data.

If you have any ideas about how to make this faster, please let me know. — fmalina, Feb 02 '11 at 19:46
you could cache the result, save the aggregates to their own model, run some sort of background aggregation, those types of things... — tijs, Aug 15 '12 at 11:44

How to aggregate barchart data for volumes of registrations per day from Django auth?

2 Answers2