Let Students Decide: Which Classes Truly Matter

• 

  Average Rating of class to determine if it should be kept for the next school year.

• 

  Overall Learning Style Preferences Barplot

• 

  Overall Work Preferences Barplot

• 

  Work Preferences for selective learning preference style Barplot

• 

  Learning Styles for selective work preference style Barplot

• 

  Raw data for enjoyable rating

• 

  Raw data for impact rating

• 

  Raw data for practical rating

Inspiration

As students part of a new high school, we saw the issues our school ran into when deciding what classes to offer the freshmen and sophomores in its first year. We wanted to create a program that would eliminate indecisiveness and help schools make data-driven decisions about the classes they can offer, especially elective classes, which there is an abundance of. Students deserve to have classes that truly matter to them and schools shouldn't have to stress about staffing or financial measures to supplement the students' growth.

What it does

The program predicts the future class of student ratings using three aspects/features: practicality, Impact, and enjoyability. Using machine learning, it creates a model (Decision Tree Regressor) that learns from past student ratings and then predicts the ratings for future students. This will help educational institutes decide whether the class will continue to be offered next year based on whether the future ratings produce an average above 2.5 out of 5.

How we built it

We built the program using Python, pandas, scikit-learn, and Decision Tree Regressor. Pandas handles the data by reading the Student Survey.csv file and prepares the features and targets for predictions. A decision tree regressor is a machine-learning model used to predict outputs. Scikit-learn trains the Decision Tree Regressor model on the specified features and predicts ratings accordingly.

Challenges we ran into

At first, we wanted to use Python Flask and APIs, but our computers kept on producing errors saying that they weren't able to find the module. That's when we switched to using a Google Colab Notebook. Additionally, the model was trained on a limited sample dataset, which most likely produced less reliable predictions and we only predicted the ratings for three categories from the dataset.

Accomplishments that we're proud of

We created a model that predicts student ratings and produces graphs that provide insight into the class's standing. The bar graphs were created based on user input with drop-down features.

What we learned

We learned how to create a machine learning model, extract data from a .csv file, and create bar graphs that can provide valuable insight in real-world scenarios.

What's next for Let Students Decide: Which Classes Truly Matter

We would like the models to handle larger datasets and incorporate more numerical and language data. Additionally, we would like to develop a more user-friendly interface by using a web platform instead of a Notebook. This could potentially allow users to easily add more data and view predictions and insights.

Built With

• googlecolab
• pandas
• python