In this guide you'll learn the key concepts of working with PYLON for LinkedIn Engagement Insights.
What is LinkedIn Engagement Insights?
LinkedIn Engagement Insights is a data source that you can analyze using PYLON. The source provides real-time access to shares, engagements, and clicks from members and companies, taking place on newsfeeds across LinkedIn.
For all activities a rich set of demographic and content fields are available, allowing you to perform detailed audience analysis.
What is PYLON for LinkedIn Engagement Insights?
PYLON is DataSift's platform which allows analysis of activities on social networks using DataSift's privacy-first approach.
PYLON for LinkedIn enagagement insights is the product you will use to analyze LinkedIn activity. It is an instance of PYLON that is installed within the LinkedIn firewall which:
- ingests real-time data feeds from LinkedIn.
- records activities and events into an index ready for analysis.
- allows you to run analysis queries against the recorded activities and events.
- respects member privacy by returning aggregated results, giving audience-level insights without providing any private data.
What is the LinkedIn 'shared index'?
All activities and events fed to PYLON by LinkedIn are called interactions. These interactions are recorded to an index ready for analysis.
In PYLON for LinkedIn Engagement Insights all interactions are recorded to a single index, called the shared index. It is called the shared index because all product customers have access to the same index. This recording is always running, providing the most recent 30 days of interactions for analysis. It is this shared index that you will analyze.
What data is included?
LinkedIn Engagement Insights covers the following types of interactions on LinkedIn:
- Members and companies sharing posts publicly.
- Members and companies liking and commenting on posts, and resharing posts publicly.
- Members and companies clicking on shares that appear in their newsfeed.
- Members and companies clicking on articles authored by members on LinkedIn, or shared from other websites.
Activities and events from all countries are included in the 30-day recording.
The LinkedIn data model
The rest of the developer guide covers the data model in detail, but let's quickly introduce the basic concepts you'll need to understand.
Interactions are generated by members and companies on LinkedIn. They relate to the activity being carried out and are divided into two types: activities and events.
Activities are social actions carried out by members and companies, including the shares they post, and the engagement on these shares. Events are similar to tracking events you may have seen in web analytics platforms, and include clicks and page views.
|Members and companies||
Interactions are triggered by two types of actors; members and companies.
The people who log in to LinkedIn to share and engage with content are called members. Companies can create a presence on LinkedIn by creating a company page. When a company page administrator comments or likes a company update, it appears as the company commenting or liking the company update.
The data model provides a wide range of demographic details for members, including their industry, seniority level, country and which companies they follow.
A piece of content that is shared (via a url) is called an article.
Article details include the title, summary, domain and url of the content, all of which you can use in your analysis.
Named entities are extracted from the articles that are shared or engaged with. These entities allow you to analyze what topics and entities are popular.
Named entities include brands, well-known people, concepts and locations.
Companies that are mentioned in posts and articles are extracted for analysis.
How do you work with PYLON for LinkedIn Engagement Insights?
Unlike other DataSift products you do not need to specify the interactions you would like to record for analysis. Instead you analyze the shared index which is always populated with all activities and events from the last 30 days.
This allows you to focus on performing your analysis. You can quickly test and iterate your queries against the latest data at any time.
Running analysis tasks
You run your analysis tasks using the aynchronous Task API. For each analysis query you will:
- submit a new task of type 'analysis'.
- wait for the analysis task to complete.
- fetch the analysis task results.
The Task API is used because some analysis tasks are complex and are run against an index containing a large quantity of interactions, so therefore can take a few seconds to complete.
As you get more familiar with the Task API, you'll see that you can submit batches of queries, retrieve the results of the batch, and build complex reports.
The next guide looks at the activities and events that take place on LinkedIn which you can access in your analysis.