A key challenge in big learning analytics is how to aggregate and integrate raw data from

multiple, heterogeneous sources, often available in different formats, to create a useful educational data set that reflects the distributed activities of the learner; thus leading to more precise and solid LA results. Furthermore, handling of big data is a technical challenge because

efficient analytics methods and tools have to be implemented to deliver meaningful results without too much delay, so that stakeholders have the opportunity to act on newly gained information in time.

Learning is increasingly taking place in open and networked learning environments. In the past few years, Massive Open Online Courses (MOOCs) have gained popularity as a new form of open learning. MOOCs can be roughly classified in two groups. On the one hand there are xMOOCs. Although they gained a lot of attention they can be seen as a replication of traditional learning management systems (LMS) at a larger scale. Still they are closed, centralized, structured, and teacher-centered courses that emphasize video lectures and assignments. In xMOOCs all services available are predetermined and offered within the platform itself. On the other hand there is the contrasting idea of cMOOCs combining MOOCs with the concept of personal learning environment (PLE). In contrast to xMOOCs, cMOOCs are open-ended, distributed, networked, and learner-directed learning environments where the learning services are not predetermined, and most activities take place outside the platform (Yousef et al., 2014a)

https://www.researchgate.net/publication/278712499_Learning_Analytics_Challenges_and_Future_Research_Directions

Learning analytics have the power to enhance eLearning experience and create more effective eLearning environments:

Helps to predict learners’ performance
One of the most significant benefits of analytics is that they can provide insight into not only how a learner is performing today, but also about his/her future performance throughout the duration of the eLearning course. For example, online facilitators may foresee if a particular learner is likely not to pass the eLearning course, or if the learner is likely to pass the eLearning course if additional support is provided (such as further readings or tutoring sessions). Therefore, analytics can help to determine if learners may benefit from supplementary eLearning materials and/or peer/instructor aid throughout the eLearning course. This leads to higher grades and a more meaningful and complete eLearning experience.
Provides learners with a personalized eLearning experience
Through learning analytics, eLearning professionals and online instructors gain the ability to custom tailor eLearning experiences for each and every individual learner. If the data shows that a learner is taking a great deal of time to finish a particular eLearning module, then appropriate measures can be taken to offer learner more customized educational tools and eLearning course resources. For example, learners can be provided with links to sites that may help them to effectively comprehend the topic, or videos that allow them to learn through a more auditory/visual approach. No two learners are alike, and learning analytics gives eLearning professionals the power to ensure that no two eLearning experiences are alike either.
Increased learners’ retention rates
Given that more learners have the opportunity to enhance their performance thanks to learning analytics data and intervention; fewer learners will drop out or fail the eLearning course. If a learner isn't faring well throughout the eLearning course, then he/she is less likely to be motivated to remain enrolled. As a result, a learner will simply stop participating, which means that institutions and/or organizations may see a steep decrease dropout rate and/or profits and learners simply won't benefit from the informative eLearning courses that are being provided.
Helps to improve future eLearning courses
Not only can learning analytics help current learners, but can also help future learners as well. For instance, if the data shows that a vast majority of learners are finding one particular aspect of the eLearning course too challenging, then the developers can change the difficulty level of that specific eLearning module. This will lead to more powerful and impactful eLearning environments tomorrow, thanks to the data that has been collected today.
Boost in cost efficiency
If you can gain an in depth understanding of how the eLearning courses and their respective resources are being utilized, and how learners are actually acquiring information, and which aspects of the deliverable are successful (and which are falling short, for that matter), then you have the power to achieve higher quality eLearning at a lower cost. For example, if you determine, through analytics, that a particular section of the eLearning course simply isn't helping learners to achieve their learning goals, then you can devote your resources to either improving it or focus on another area that may be a more worthwhile investment.

https://elearningindustry.com/5-reasons-why-learning-analytics-are-important-for-elearning

A number of challenges faced are:

Challenge 1: Unit design. As the VLE (virtual learning environment) is usually the core system from which most of the LA data is sourced, the first challenge is that limited use of the VLE and its tools will lead to the collection of less data which will have a negative effect on the quality of the predictions generated by the LA engine.

Challenge 2. Co-ordination across the institution. The data for inclusion in LA can be sourced from many locations. The co-ordination challenge is the process of managing the myriad of potentially useful information to generate a consistent whole that efficiently and effectively supports student learning.

Challenge 3. Time. In order for the LA engine to be able to provide predictions there needs to be data available from past iterations of a unit. Over time and with more data to include, the predictive engine can become more accurate. There is a double challenge here in that the need for consistent and voluminous data would suggest that the unit should remain unchanged whilst LA is being used and also that it could take some years before accurate predictions are available.

https://microsites.bournemouth.ac.uk/flie/2018/05/14/the-challenges-of-learning-analytics-and-possible-solutions/