Storing and processing raw data from myTracker

News feed

Data is a highly valuable resource, especially in marketing. More data means more precise analysis and better business processes. This article will focus on how to best store and process myTracker data.

No limits

Architecture-wise, almost any tool will do, be it column-oriented databases, hardoop, or internal BI. Data export destination depends on your tech stack. If we take the generally accepted approach to data storage, column-oriented databases are the most popular solution:

  • They provide for an easier way to store large amounts of data;
  • Partitioning (e.g. by date) enables faster processing;
  • Advanced interfaces for data aggregation and analytics;
  • Since processed data is static, there is no need for UPDATE and DELETE options.

There are no limitations – myTracker exports data to JSON or XML files, which can be stored anywhere. However, it is important what data you export, because it takes up the usually scarce space. So it would make sense to select only the key metrics required for the current purposes to be exported and leave the rest, as myTracker has no limit on data storage time – records for any period are always at your fingertips.

How to obtain information quickly

You can set up regular exports of required data from myTracker through Export API. For instance, you can sort by devices that have reached a particular goal, or by downloads – the data pool is vast and can be filtered when creating a request for information, e.g. on a specific campaign or tracking link.

Filter by date of entry

For example, the user created an account on 1 January but has no internet access until 20 May. In this case you receive a sign-up notification on 20 May instead of 1 January. The event will not be discarded – myTracker uses the very date when the event occurred.

AvailabilityTsFrom and AvailabilityTsTo filters provide for the output of recent data that you have not exported previously. Imagine you have 1,000,000 events dated 1 January that you have already processed, and then you add another 1,000. These options allow you to set aside the events you have processed and focus on the new 1,000. As you can see from the chart below, if the user exports data on 3 January, instead of processing all 3,000,000 events, they can select the new ones only – 1,000 for 1 January and 3,000 for 2 January.

This way myTracker allows you to quickly obtain the data you need right now, at the same time making sure that any other information remains available and does not expire. Our earlier article on exporting raw data from myTracker is available here.