Analysis of user behavior based on data from PANORAMA

Our multidisciplinary team works to produce and implement the best e-learning solutions, adapted to the specific needs of each organization.

Among the services we offer, a very important one is Usage and rating reports, which allows us to analyze the performance of each course and each user/student with a detailed report. Is considered date to information stored in some digital format that can later be used as a basis for analysis and decision making. This information is collect, organize and analyze to investigate various issues: data analysis is a process that consists of inspecting and transforming it, with the aim of highlighting useful information to suggest conclusions and support decision-making. Once this happens and we have the reports ready, we are in a position to predict future situations.

To address this diverse spectrum of analysis, en Aulasneo we have developed PANORAMA: a flexible analysis engine that can be easily customized and adapted to the specific needs of each organization that delivers online courses through Open edX ™.

PANORAMA is an analysis engine that allows organizations that use Open edX ™ to offer online courses

What is the importance of the data extracted from PANORAMA?

As a first particular case, let us consider video user behavior analysis based on the data obtained. We have several possible approaches, for example analyzing the data by video or by user/student. Regarding the analysis according to video, the following questions may arise:

How many times was a video viewed?

how many times was it overlooked? (that is: when a student enrolls in a certain course, they are expected to see certain videos that are part of it, and by extracting the data we are able to know if the student actually saw the video or not)

what part of the video was seen?

how many times was it rewound?

and advance?

in which part(s)?

How long did it take, from the first “play”, to finish watching it?

from here, we can try to discover some pattern of behavior: Are the most overlooked videos the longest? And the ones that were rewound the most? Are they from the same class? And the ones that were seen the most?

Depending on the user, we can mostly measure performance: how many videos did you miss? How many did you start watching and not finish? Did you have to rewind them many times? How many total hours did he spend watching videos?

From here, we can start combining the results to generate new ones:

who/which was the most active student/video, how much difference there was with the least active, if they belong to the same course, and a thousand etcetera.

As a second particular case, let us consider the analysis of user behavior when answering questions and doing exercises. Similarly, we can ask ourselves:

how many times a question was answered correctly,

how many incorrectly and how many students answered it incorrectly (remember that a student may have answered incorrectly several times in the same question),

how many was overlooked,

What was the question that was answered the most times correctly?

which was the student who answered correctly the most times... and a thousand etceteras.

The questions we can ask ourselves are endless., but the importance of data analysis lies not only in asking several good questions that allow us to obtain relevant information, but especially in the after: how we use that information to improve the learning process for everyone

julia zack
DevOps Analyst