Machine Learning Group Project: Create and Test Your Own Model

In this group project, you’ll use machine learning skills to create your own prediction model in Google Colab. Your group will pick a fun topic you’re interested in, choose a yes/no question, build a dataset, train a model, and test it with another group’s inputs in front of the class!

1. Choose Your Question

Why does this matter? Machine learning starts with a question you want to answer, like predicting if something will happen or if someone will like something. Choosing your own topic and question makes this project creative and exciting!

Your role in this step: As a group, brainstorm a yes/no prediction question about any topic you like—sports, entertainment, nature, or anything else that sparks your interest!

2. Create Your Dataset

What’s the big picture? A dataset is a table of information with features (like sunlight hours) and an outcome (like plant growth) to train your model.

What’s your job here? Your group will invent a small dataset (10-15 rows) based on your question, making up realistic data for different scenarios, objects, or people.

1. Decide on at least 7 features that affect your prediction (e.g., for “Will a plant grow well?”: Sunlight Hours, Water Amount, Soil Type).
2. Create a table with 15 rows, each showing a scenario with a yes/no outcome (0 = No, 1 = Yes).
3. Use this code template in a new Colab notebook (name it “Group [name] ML Project”):

import pandas as pd

# Create your dataset (edit this!)
data = {
    'Feature1': [10, 8, 5, 12, 3, 9, 7, 4, 11, 6],  # e.g., Sunlight Hours
    'Feature2': [7, 6, 5, 8, 4, 7, 6, 5, 8, 6],   # e.g., Water Amount
    'Feature3': [1, 1, 0, 1, 0, 1, 0, 1, 1, 0],   # e.g., Good Soil (0=No, 1=Yes)
    'Outcome': [1, 1, 0, 1, 0, 1, 0, 1, 1, 0]     # e.g., Grows Well (0=No, 1=Yes)
}
df = pd.DataFrame(data)

# Show the dataset
df
    

3. Explore Your Data

Why is this step key? Exploring your data helps you spot patterns, like whether more sunlight helps plants grow, so you understand what your model will learn.

Your task: Summarize your dataset and plot one feature against the outcome to look for connections.

1. Add a new code cell and paste this code for statistics:

# Show summary statistics
df.describe()
    

2. Add another code cell for a scatter plot (replace Feature1 with your feature’s name):

import matplotlib.pyplot as plt

# Scatter plot
plt.scatter(df['Feature1'], df['Outcome'], color='green')
plt.xlabel('Feature1')  # e.g., Sunlight Hours
plt.ylabel('Outcome (1=Yes, 0=No)')
plt.title('Feature1 vs. Outcome')
plt.show()
    

4. Train and Test Your Model

What’s the goal? You’ll train a Decision Tree model to learn patterns in your data and test how well it predicts your outcome.

Your role in this step: Prepare your data, train the model, and check its accuracy.

1. Add a new code cell and paste this code (edit feature names to match your dataset):

from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier
from sklearn.metrics import accuracy_score

# Features and target
X = df[['Feature1', 'Feature2', 'Feature3']]  # Edit with your feature names
y = df['Outcome']

# Split into training and testing sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)

# Train the model
model = DecisionTreeClassifier(random_state=42)
model.fit(X_train, y_train)

# Predict on the test set
y_pred = model.predict(X_test)

# Show predictions and accuracy
print("Test set features:\n", X_test)
print("Predictions:", y_pred)
print("Actual:", y_test.values)
accuracy = accuracy_score(y_test, y_pred)
print("Model accuracy:", accuracy * 100, "%")
    

5. Prepare to Test with Another Group

What’s exciting about this? You’ll share your model by predicting outcomes for another group’s inputs.

Your task: Write code to take inputs from another group and plan your class presentation.

1. Add a new code cell and paste this code (edit feature names and prompts):

# Get inputs from another group
print("Enter details for prediction:")
feature1 = float(input("Feature1 (e.g., Sunlight Hours): "))
feature2 = float(input("Feature2 (e.g., Water Amount): "))
feature3 = float(input("Feature3 (e.g., Good Soil, 0=No, 1=Yes): "))

# Create new data
new_person = pd.DataFrame({
    'Feature1': [feature1],
    'Feature2': [feature2],
    'Feature3': [feature3]
})

# Predict
prediction = model.predict(new_person)
print("Prediction:", "Yes" if prediction[0] == 1 else "No")  # Edit "Yes/No" to match your question
    

6. Present and Reflect

Why present your work? Sharing your model helps you explain what you learned and see how others tackled their projects, like scientists sharing discoveries.

What are you doing? Present your model, test it with another group’s inputs, and reflect on your project to deepen your understanding of machine learning.

1. Present to the class:
   • Explain your question and why you chose it.
   • Show your dataset (run df in Colab).
   • Share your model’s accuracy and any patterns you found.
   • Ask another group for inputs, run the prediction code, and discuss the result.
2. Add a text cell in Colab or write on paper to reflect:
   • What patterns did you find in your data?
   • What was the hardest part of this project? How did your group solve it?