fs-backend.md

Context: At Clarisights, we pull data from a lot of different APIs (for eg. fb, adwords, twitter, adjust etc) and store it in our database system. Consider a scenario where we are pulling data from Google Ads API. The data that we pull consists of different dimensions and metrics. It looks something like the following -

Query	Count
Best web hosting	1000
Cheap web hosting	410
Go daddy web hosting	20
Other web hosting	40

The above data represents the number of times (Count), each query was queried. In this case, the possible values for Query is infinite. Marketers usually do not care about queries which do not have significant count value.

Problem Statement:

1. We run a job which pulls data from the API everyday. The API returns a CSV file as response.
2. The API can take parameters - start date, end date, {list of dimensions}, {list of metrics}
3. Marketers do not care about individual queries which have​ < 50 Count till date or < 50 Count in a month.
4. We need to still show these queries, but all of them are grouped together as a single query Queries with Count < 50.

As an example, for the above API response the data we show on our platform can be something like -

Query	Count
Best web hosting	1000
Cheap web hosting	410
Queries with Count < 50	60

Expectations - You need to come up with a solution as to how we can process and store the data. You have the freedom to use any language/framework to approach this problem statement. Things we specifically tend to look for:

1. Processing Logic
2. Code Quality
3. Test Coverage

Here's a sample input for you - https://clarisights-users.s3.eu-central-1.amazonaws.com/assignment/sample+data.csv