OgiBalboa/sample-api

Adjust Performance Metric Monitoring

About

This API is an interface for a sample dataset. You can retrieve a single row or list rows by query parameters.

Installation

Python 3.8 or newer is required.

pip install -r requirements.txt

To run

python manage.py runserver

Usage

You can filter, order, group, aggregate data by using query parameters with single end point.

http://localhost:8000/api/v1/performance_metrics/

/api/v1/performance_metrics is the only end point you need to perform complex queries on performance_metrics table.

Query Parameter Structure

Selecting Columns

You can select the columns/fields that you want to fetch from db by using query parameter "column". You should use query parameter multiple times to assign multiple parameters to it.
Example

http://localhost:8000/api/v1/performance_metrics/?column=channel&column=os&column=date

Filter / Search

Columns of dataset table are : date | channel | country | os | impressions | clicks | installs | spend | revenue
You can filter any of these columns by using django orm lookups (guide link: https://docs.djangoproject.com/en/dev/ref/models/querysets/#field-lookups)
Example

http://localhost:8000/api/v1/performance_metrics/?os__in=ios,android
http://localhost:8000/api/v1/performance_metrics/?clicks__gte=1000

Ordering

You can order data by any column (this could be limited by columns like clicks, installs, spend and revenue) by declaring ordering parameter in query params. For descending order add (-) to field.
Example

http://localhost:8000/api/v1/performance_metrics/?ordering=-clicks

AGGREGATION

You can use aggregation functions like Sum and Count by passing query parameter (sum=(field), count=(field)).

Grouping

You can group your data by adding group_by query param. You can use this parameter multiple times to group 1 or more fields.
Example

http://localhost:8000/api/v1/performance_metrics/?ordering=-click&sum=clicks&date__lte=2017-06-01&group_by=clicks&sum=impressions&group_by=channel&group_by=country

Special Cases

CPI Metric

You can get cpi (cost per install) metric by passing get_cpi=True query param.

Date Ranges

You can specify start_date as date__gte and end date as date__lte

Common API use-cases

1. To Show the number of impressions and clicks that occurred before the 1st of June 2017, broken down by channel and country, sorted by clicks in descending order

http://localhost:8000/api/v1/performance_metrics/?ordering=-clicks&group_by=channel&group_by=country&date__lte=2017-06-01&column=clicks&sum=clicks&column=impressions&sum=impressions

2. To Show the number of installs that occurred in May of 2017 on iOS, broken down by date, sorted by date in ascending order.

http://localhost:8000/api/v1/performance_metrics/?ordering=date&group_by=date&sum=installs&date__gte=2017-05-01&os=ios&date__lte=2017-05-31&column=os

3. To Show revenue, earned on June 1, 2017 in US, broken down by operating system and sorted by revenue in descending order.

http://localhost:8000/api/v1/performance_metrics/?ordering=-revenue&group_by=revenue&date=2017-06-01&group_by=os

4. To Show CPI and spend for Canada (CA) broken down by channel ordered by CPI in descending order:

http://localhost:8000/api/v1/performance_metrics/?ordering=-cpi&get_cpi=True&column=channel&country=CA&column=country

TO DO's

• Adding a validator for query params required (to prevent data leakage)
• Tests should be written.
• Authentication is a must.
• Throttling could be helpful to secure API.
• API returns the field you aggregate even you did not specify in column params.