Analysis of scores from three different exams and a variety of personal, social, and economic factors that have interaction effects upon them. The goal is to see which of those factors contribute to scoring better/worse in exams. This could give students and individual schools helpful insights. In the end, I also built a classifier that predicts whether a student will score above or below average based on certain factors.
pandas, numpy, matplotlib, seaborn
Initial Exploratory Data Analysis and Preprocessing
- Display of basic dataset statistics
- visualizing average exam distribution
- preprocessing the data for in depth analysis of the following questions
Analysis
-
Do students with parents that have a college degree perform better?
-
Does having a free/reduced meal have an impact on the performance?
-
The Gender race: Who scores better in which subject? Who scores better in total?
-
Does having a good score in one subject contribute to having a good score in another subject?
-
Do race differences show different performances?
-
Do people that complete the test preparation course score higher?
Predict performance
- SVM to predict above or below average performance
Conclusion
- Breaking down the results for the questions in the Analysis
Next Steps 27.09.2019
- Add more features/data and incraese performance of the model
- Currently working on a interative dash-plotly dashboard where a user can input the different factors discussed in the project and get the performance of the student as output from the SVM I built in the end of the project.
Progress:
- User can Input the different features, I recieve them with linked callbacks, preprocess them and fit them into the SVM.
ToDo:
- Front-end
- Plotly Plot for Model Output probabilities (above/below average performance probabilites).
