Classification of Test Scores

Introduction to the Problem I am interested in seeing if parents' children's test scores have any correlation to the parent's level of education. I will look deeper into seeing if you can classify a student's test scores to get their parent's level of education.

About my Dataset The dataset being used for this project is called Students Performance in Exams found on Kaggle. Click here for the link to the website page. The following dataset gives information about a person's test scores in Math, Reading, Writing, and their parent's level of education.

Preprocessing

I decided instead of using 3 separate values for the Math, Reading, and Writing scores I would combine these values into a single number. In the excel file, I used a command to add the values of the 3 columns for each person. I also added an average variable of the total score as another column. Since the 3 values are now unnecessary I went ahead and dropped them from the dataset.

I then checked the data types to make sure I was working with Integers and not Strings. Since some variables were Strings I converted them to Integers using get_dummies().

Visualization

Shown below we can see a typical bell curve distribution representing a normal distribution.

Modeling & Evaluation

I used a decision tree to help with my classification.

I also attempted to use logistic regression to evaluate my data based on the reason my dataset was using 0 and 1. The data value was 0 if the parent didn't have the sorted education level or 1 if they did have the specified education level.

Story Telling

Based on the two models you are able to see a negative correlation for students whose parents have some high school education. Students with parents that have a bachelor's degree look to have a positive correlation. Using this information I am able to tell that a parent's educational level does play a small role in the child's education.

Impact

This conclusion shows a negative impact on families with low educational levels. It may not just be the overall fact that the parents have less education but rather whether or not they take the time to help their children learn subjects they are struggling with. Parents that did not take their own time when they were learning prior may not see fit to help their children as well. The learning environment of children is very important to promote increased learning.