Getting Started

I have decided to defined 2 endpoints, mostly for the different response types, 'totals' with just a number and the other endpoint that returns a list of items.

CTR has been considered a metric in order to simplify the number of endpoints.

Reference Documentation

There are the following metrics:

• CTR (requires a groupBy params)
• Clicks
• Impressions

There are the following aggregations:

• SUM
• AVG
• MAX
• MIN

All endpoints can be filtered by :

• Daily ("MM/dd/yyyy"): Filters by a given date
• DateFrom ("MM/dd/yyyy"): Filters by date greater or equals than the given dateFrom
• DateTo ("MM/dd/yyyy"): Filters by date lower or equals than the given dateTo
• DateFrom and DateTo: Filters by the given date range
• Datasource: Filters by datasource value
• Campaign: Filters by campaign value

The endpoints that support grouping allow the following fields to be grouped by:

• daily
• campaign
• datasource

There are 3 endpoints:

• /{metrics}/total/{aggregations} Calculates the total for the given metric and the given aggregation. It does not support grouping
• /{metrics}/{aggregations} Calculates the aggregation of the given metric. Supports grouping.
• /{metrics} Calculates the sum of the given metric. Supports grouping. This is the endpoint to use when using CTR because adding an aggregation on top of the CTR is not supported.

Application deployment

The application has been deployed in this environment:

http://simpledatawarehouse-env.eba-v3pxyp9n.eu-west-1.elasticbeanstalk.com/marketing/

Examples

• Will return total clicks for a given Datasource for a given Date range

curl --location --request POST 'http://simpledatawarehouse-env.eba-v3pxyp9n.eu-west-1.elasticbeanstalk.com/marketing/clicks/total/sum' \
--header 'Content-Type: application/json' \
--data-raw '{"datasource": "Google Ads","dateFrom":"09/12/2019","dateTo": "11/12/2019"}'

• Will return click-Through Rate (CTR) per Datasource and Campaign

curl --location --request POST 'http://simpledatawarehouse-env.eba-v3pxyp9n.eu-west-1.elasticbeanstalk.com/marketing/ctr' \
--header 'Content-Type: application/json' \
--data-raw '{"groupBy": "datasource,campaign"}'

• Will return impressions over time (daily)

curl --location --request POST 'http://simpledatawarehouse-env.eba-v3pxyp9n.eu-west-1.elasticbeanstalk.com/marketing/impressions/sum' \
--header 'Content-Type: application/json' \
--data-raw '{"groupBy": "daily"}'

Possible improvements

• Add a way to combine CTR with some aggregations, like for example: return the AVG(CTR) for a month
• Add Pagination, so if there is a lot of data you can consume it in pages.
• Add better sorting. At the moment is asc by default and using the values from the groupBy param.
• The way how JPA Criteria builds the custom objects, seems quite messy or at least I couldn't find a better way.It expects a certain constructor with the right param types and in the right order, rather than using the default constructor and use the setters for setting the values. Having a dynamic groupBy forces you to add many constructors with all order and fields possible.