The more learning data you have (assuming it’s good data), translates into better, more comprehensive reporting and insights around your training programs and initiatives. For instance, you can create reports from multiple source systems, import it into a BI tool or data warehouse, or merge training data with your HR data so you can slice and dice that info any way you want. But exporting learning data to a BI tool isn’t always easy. As a result, sharing important L&D metrics across the business can be quite hard. In this blog post, we’ll cover the best ways to extract that data, including how to get around some common barriers that may arise during the process.
What’s the best method for extracting data from an LMS or LXP?
The answer depends on the actual tool or platform that’s being used. For instance it might be xAPI, API, or a combination of both. In most scenarios, you may need to work with the learning record store (LRS) and/or platform vendor to find the solution that best fits your needs.
Generally speaking, though, the first extraction method is xAPI. (For example, Watershed has a way to extract data from pretty much any LRS, even if it's not fully conformant. And many LRSs have some xAPI functionality that can push xAPI data to Watershed as an activity provider.)
The second method is via an API connection. Most platforms have APIs that allow access to its data or let you generate and download reports. Accessing APIs does require some technical skill and understanding, the relevant documentation, and the correct credentials and tools.
And the last method is via CSV/Excel file. Again, most platforms have some way for you to download a report allowing you to play with the data in Excel or the data tool of your choice. Many BI tools allow you to upload CSV data for further analysis in combination with other data.
What are some challenges when it comes to extracting or moving data?
Simply put, this process can be difficult and time consuming. It requires people with skills in ETL—or extract, transform and load—the tools and software required, and lastly somewhere for the data to be stored and analyzed. This can be as simple as an Excel file, or as complicated as a data lake or BI tool.
Extracting data usually shows all the problems with your data. Our experience when looking at our customers' data is it's not until you begin to download, view, and try to merge your learning data with other datasets that you find out the issues with it. At best this can be simple things that can easily be overcome like inconsistent usernames, and at worst, large amounts of it are simply missing or intelligible!
People and systems can be quite precious about their data. And, often, this results in senior-level employees having to tell those people they need to help and provide the needed data.
So how do you overcome these data extraction challenges?
The first step to overcoming data extraction challenges is simply knowing that they exist. And as a result, you can focus on identifying:
- whether the data is extractable via CSV, xAPI, or API;
- who can help you get access to that extraction method; and,
- what errors lie in your data.
Up Next: How does an LRS fit into a data ecosystem?
In the final blog post in this series, we’ll discuss how a learning record store fits into the larger data ecosystem and how your learning data can mix and mingle with other organizational data.
About the author
Subscribe to our blog