The initial step necessary to be able to recommend any content based on users' behaviour is the data acquisition on websites and applications. This includes page views of articles, playback of audios and videos, reading of articles, selection of a program, and more.
PEACH provides libraries in order to easily and consistently collect data on all platforms (web sites, iOS mobile apps and AppleTV, Android mobile apps, AndroidTV).
It is important to note that the data should be collected on a maximum possible outlets where the content is consumed in order to improve the quality of the recommendation.
To not hesitate to get in touch with us so that we can establish together which platform / product will provide the best scenario for a proof of concept, and will provide good data to start your journey.
Our open source data collection libraries
- Android : https://github.com/ebu/peach-collector-android/
- iOS : https://github.com/ebu/peach-collector-ios/
- web : https://git.ebu.io/pipe/peach-collector-js
- Common data model
- Helper functions to collect typical events
- Logic to automatically extract technical details of the client
- Support both anonymous and logged-in users
- Local storage of events when app is offline
- Possibility to store locally events, and send them in batch
- Events can be sent to PEACH and other data consumers
- Client identifier can be share across multiple apps of the same provider
PEACH and Tag management systems
PEACH can receive data through the use of a Tag Management System (TMS). Some work client-side (dynamically injecting js code in your pages to send data on relevant pages / actions), other server-side (a proxy distributes the data to multiple data sinks) or both. We had some good experience with TMS to collect data to PEACH and some bad ones, depending on the flexibility of the system.
Batch data delivery
In case you already have a data collection pipeline in place, PEACH can also receive / fetch data in batch. This is sometimes easier to get started with PEACH.
However, direct data collection to PEACH is the preferred route for a mature integration. We have dedicated realtime pipeline that will only work when data is sent directly to PEACH from the client. For example, we use it to remove from recommendation lists the elements you have just watched.