The dataset used in this report contains information on user activity in sharing and reading articles. This data is taken from Deskdrop, an internal article sharing platform for the CI&T team. This dataset contains information about users, articles shared and read, and activity times.
By using 2 models, namely Content Based Filtering and Collaborative Filtering and in the evaluation stage used Top-N accuracy metrics. From the evaluation results in getting better model results is Collaborative Filtering. And at the testing stage, the results are quite accurate on Top N article recommendations.