Abstract:

Data mining has become popular in the last few years. Its importance has grown so much that it is now considered a competitive advantage by some companies. Data contains a lot of useful insights and if they are harvested, it can improve the processes many times.

One problem commonly faced is that the insights or the results harvested from data through data mining are too complex or messy to be understood by a layman decision-maker who is simply illiterate about data mining procedures. Another issue is that the decision-makers get their hands on the generated insights only and can’t play with the data because their experience with coding is almost non-existent.

In this project, what we are going to do is that we have got our hands on a dataset of Portuguese schools. The dataset is comprehensive. It has got many columns about the students’ biodata. And finally, it has 3 columns with the grades achieved by students in the exams. With this dataset, we can generate some good insights to see the most important factors which contribute to students’ success so the school and professors can use the available resources effectively to optimize them.

First of all, data is cleaned and it is made sure that it is good enough for making some plots and doing some exploratory analysis. After that, we make useful and informative charts about the data which help understand the general trend in the data. After that, we use machine learning algorithms to predict the final grade of the students. Once we have achieved a good accuracy score with an algorithm, we make a very interactive, comprehensive, and informative dashboard about the data representing all the necessary stuff we generated in the previous steps. The dashboard would also give the option to the user/stakeholder to put custom values and get the predicted grade.

In short, our purpose is to generate useful insights, implement a prediction algorithm and make it easy for the stakeholders to interact with data and get the maximum benefit from it by trying custom values as the input to get a predicted grade.

Keywords:

Data analysis, secondary school student performance prediction, dashboard, learning analytics.