diff --git a/.DS_Store b/.DS_Store
new file mode 100644
index 0000000..5008ddf
Binary files /dev/null and b/.DS_Store differ
diff --git a/.history/Analyze_20240207183142.py b/.history/Analyze_20240207183142.py
new file mode 100644
index 0000000..aa9943c
--- /dev/null
+++ b/.history/Analyze_20240207183142.py
@@ -0,0 +1,37 @@
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def import_data():
+    """
+    Import data from the TTC Bus Delay dataset
+    """
+    return pd.read_csv('ttc-bus-delay-data-2023.csv')
+
+
+def compute_avg_delay(data, group_by: str = 'Route'):
+    """
+    Compute the average delay aggregated by specified column
+    """
+<<<<<<< HEAD
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+=======
+    return data.groupby(group_by).mean().reset_index()
+>>>>>>> origin/feature1
+
+
+def main():
+    """
+    Main function
+    """
+    data = import_data()
+    avg_delay = compute_avg_delay(data, group_by='Route')
+
+    plt.hist(avg_delay['Min Delay'], bins=30)
+    plt.title('Average Delay Distribution')
+    plt.xlabel('Average Delay (minutes)')
+
+
+if __name__ == '__main__':
+    main()
diff --git a/.history/Analyze_20240207183321.py b/.history/Analyze_20240207183321.py
new file mode 100644
index 0000000..5c23c2a
--- /dev/null
+++ b/.history/Analyze_20240207183321.py
@@ -0,0 +1,38 @@
+
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def import_data():
+    """
+    Import data from the TTC Bus Delay dataset
+    """
+    return pd.read_csv('ttc-bus-delay-data-2023.csv')
+
+
+def compute_avg_delay(data, group_by: str = 'Route'):
+    """
+    Compute the average delay aggregated by specified column
+    """
+<<<<<<< HEAD
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+=======
+    return data.groupby(group_by).mean().reset_index()
+>>>>>>> origin/feature1
+
+
+def main():
+    """
+    Main function
+    """
+    data = import_data()
+    avg_delay = compute_avg_delay(data, group_by='Route')
+
+    plt.hist(avg_delay['Min Delay'], bins=30)
+    plt.title('Average Delay Distribution')
+    plt.xlabel('Average Delay (minutes)')
+
+
+if __name__ == '__main__':
+    main()
diff --git a/.history/Analyze_20240207183332.py b/.history/Analyze_20240207183332.py
new file mode 100644
index 0000000..ff6fe76
--- /dev/null
+++ b/.history/Analyze_20240207183332.py
@@ -0,0 +1,38 @@
+#RESOLVED
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def import_data():
+    """
+    Import data from the TTC Bus Delay dataset
+    """
+    return pd.read_csv('ttc-bus-delay-data-2023.csv')
+
+
+def compute_avg_delay(data, group_by: str = 'Route'):
+    """
+    Compute the average delay aggregated by specified column
+    """
+<<<<<<< HEAD
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+=======
+    return data.groupby(group_by).mean().reset_index()
+>>>>>>> origin/feature1
+
+
+def main():
+    """
+    Main function
+    """
+    data = import_data()
+    avg_delay = compute_avg_delay(data, group_by='Route')
+
+    plt.hist(avg_delay['Min Delay'], bins=30)
+    plt.title('Average Delay Distribution')
+    plt.xlabel('Average Delay (minutes)')
+
+
+if __name__ == '__main__':
+    main()
diff --git a/.history/Analyze_20240207183335.py b/.history/Analyze_20240207183335.py
new file mode 100644
index 0000000..a0b5ad0
--- /dev/null
+++ b/.history/Analyze_20240207183335.py
@@ -0,0 +1,38 @@
+#RESOLVED FIXED 
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def import_data():
+    """
+    Import data from the TTC Bus Delay dataset
+    """
+    return pd.read_csv('ttc-bus-delay-data-2023.csv')
+
+
+def compute_avg_delay(data, group_by: str = 'Route'):
+    """
+    Compute the average delay aggregated by specified column
+    """
+<<<<<<< HEAD
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+=======
+    return data.groupby(group_by).mean().reset_index()
+>>>>>>> origin/feature1
+
+
+def main():
+    """
+    Main function
+    """
+    data = import_data()
+    avg_delay = compute_avg_delay(data, group_by='Route')
+
+    plt.hist(avg_delay['Min Delay'], bins=30)
+    plt.title('Average Delay Distribution')
+    plt.xlabel('Average Delay (minutes)')
+
+
+if __name__ == '__main__':
+    main()
diff --git a/.history/Analyze_20240211193044.py b/.history/Analyze_20240211193044.py
new file mode 100644
index 0000000..3f07dbf
--- /dev/null
+++ b/.history/Analyze_20240211193044.py
@@ -0,0 +1,33 @@
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def import_data():
+    """
+    Import data from the TTC Bus Delay dataset
+    """
+    return pd.read_csv('ttc-bus-delay-data-2023.csv')
+
+
+def compute_avg_delay(data):
+    """
+    Compute the average delay for each route
+    """
+    return data.groupby('Route').mean().reset_index()
+
+
+def main():
+    """
+    Main function
+    """
+    data = import_data()
+    avg_delay = compute_avg_delay(data)
+
+    plt.hist(avg_delay['Min Delay'], bins=30)
+    plt.title('Average Delay Distribution')
+    plt.xlabel('Average Delay (minutes)')
+
+
+if __name__ == '__main__':
+    main()
diff --git a/.history/Analyze_20240211195121.py b/.history/Analyze_20240211195121.py
new file mode 100644
index 0000000..81e17fb
--- /dev/null
+++ b/.history/Analyze_20240211195121.py
@@ -0,0 +1,34 @@
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def import_data():
+    """
+    Import data from the TTC Bus Delay dataset
+    """
+    return pd.read_csv('ttc-bus-delay-data-2023.csv')
+
+
+def compute_avg_delay(data):
+    """
+    Compute the average delay for each route
+    """
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+
+
+def main():
+    """
+    Main function
+    """
+    data = import_data()
+    avg_delay = compute_avg_delay(data)
+
+    plt.hist(avg_delay['Min Delay'], bins=30)
+    plt.title('Average Delay Distribution')
+    plt.xlabel('Average Delay (minutes)')
+    plt.show()
+
+
+if __name__ == '__main__':
+    main()
diff --git a/.history/Analyze_20240211195122.py b/.history/Analyze_20240211195122.py
new file mode 100644
index 0000000..81e17fb
--- /dev/null
+++ b/.history/Analyze_20240211195122.py
@@ -0,0 +1,34 @@
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def import_data():
+    """
+    Import data from the TTC Bus Delay dataset
+    """
+    return pd.read_csv('ttc-bus-delay-data-2023.csv')
+
+
+def compute_avg_delay(data):
+    """
+    Compute the average delay for each route
+    """
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+
+
+def main():
+    """
+    Main function
+    """
+    data = import_data()
+    avg_delay = compute_avg_delay(data)
+
+    plt.hist(avg_delay['Min Delay'], bins=30)
+    plt.title('Average Delay Distribution')
+    plt.xlabel('Average Delay (minutes)')
+    plt.show()
+
+
+if __name__ == '__main__':
+    main()
diff --git a/.history/Analyze_20240211195123.py b/.history/Analyze_20240211195123.py
new file mode 100644
index 0000000..81e17fb
--- /dev/null
+++ b/.history/Analyze_20240211195123.py
@@ -0,0 +1,34 @@
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def import_data():
+    """
+    Import data from the TTC Bus Delay dataset
+    """
+    return pd.read_csv('ttc-bus-delay-data-2023.csv')
+
+
+def compute_avg_delay(data):
+    """
+    Compute the average delay for each route
+    """
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+
+
+def main():
+    """
+    Main function
+    """
+    data = import_data()
+    avg_delay = compute_avg_delay(data)
+
+    plt.hist(avg_delay['Min Delay'], bins=30)
+    plt.title('Average Delay Distribution')
+    plt.xlabel('Average Delay (minutes)')
+    plt.show()
+
+
+if __name__ == '__main__':
+    main()
diff --git a/.history/Analyze_20240211195125.py b/.history/Analyze_20240211195125.py
new file mode 100644
index 0000000..81e17fb
--- /dev/null
+++ b/.history/Analyze_20240211195125.py
@@ -0,0 +1,34 @@
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def import_data():
+    """
+    Import data from the TTC Bus Delay dataset
+    """
+    return pd.read_csv('ttc-bus-delay-data-2023.csv')
+
+
+def compute_avg_delay(data):
+    """
+    Compute the average delay for each route
+    """
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+
+
+def main():
+    """
+    Main function
+    """
+    data = import_data()
+    avg_delay = compute_avg_delay(data)
+
+    plt.hist(avg_delay['Min Delay'], bins=30)
+    plt.title('Average Delay Distribution')
+    plt.xlabel('Average Delay (minutes)')
+    plt.show()
+
+
+if __name__ == '__main__':
+    main()
diff --git a/.history/Analyze_20240211225832.py b/.history/Analyze_20240211225832.py
new file mode 100644
index 0000000..ba0805a
--- /dev/null
+++ b/.history/Analyze_20240211225832.py
@@ -0,0 +1,43 @@
+# FIXED AND TESTED PYTHON CODE
+import pandas as pd
+import matplotlib.pyplot as plt
+import numpy as np
+
+
+def import_data():
+    """
+    Import data from the TTC Bus Delay dataset
+    """
+    return pd.read_csv('ttc-bus-delay-data-2023.csv')
+
+
+def compute_avg_delay(data, group_by: str = 'Route'):
+    """
+    Compute the average delay aggregated by specified column
+    """
+
+
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+
+    return data.groupby(group_by).mean().reset_index()
+
+
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+
+
+
+def main():
+    """
+    Main function
+    """
+    data = import_data()
+    avg_delay = compute_avg_delay(data, group_by='Route')
+
+    plt.hist(avg_delay['Min Delay'], bins=30)
+    plt.title('Average Delay Distribution')
+    plt.xlabel('Average Delay (minutes)')
+    plt.show()
+
+
+if __name__ == '__main__':
+    main()
diff --git a/Analyze.py b/Analyze.py
index 3f07dbf..ba0805a 100644
--- a/Analyze.py
+++ b/Analyze.py
@@ -1,3 +1,4 @@
+# FIXED AND TESTED PYTHON CODE
 import pandas as pd
 import matplotlib.pyplot as plt
 import numpy as np
@@ -10,11 +11,19 @@ def import_data():
     return pd.read_csv('ttc-bus-delay-data-2023.csv')
 
 
-def compute_avg_delay(data):
+def compute_avg_delay(data, group_by: str = 'Route'):
     """
-    Compute the average delay for each route
+    Compute the average delay aggregated by specified column
     """
-    return data.groupby('Route').mean().reset_index()
+
+
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+
+    return data.groupby(group_by).mean().reset_index()
+
+
+    return data.groupby('Route')['Min Delay'].mean().reset_index()
+
 
 
 def main():
@@ -22,11 +31,12 @@ def main():
     Main function
     """
     data = import_data()
-    avg_delay = compute_avg_delay(data)
+    avg_delay = compute_avg_delay(data, group_by='Route')
 
     plt.hist(avg_delay['Min Delay'], bins=30)
     plt.title('Average Delay Distribution')
     plt.xlabel('Average Delay (minutes)')
+    plt.show()
 
 
 if __name__ == '__main__':
diff --git a/DSI_git_assignment b/DSI_git_assignment
new file mode 160000
index 0000000..ce31e28
--- /dev/null
+++ b/DSI_git_assignment
@@ -0,0 +1 @@
+Subproject commit ce31e28393811bbc669a05dba9fd34ad8949a5a6
diff --git a/README.md b/README.md
index c9d8518..0a1dcf1 100644
--- a/README.md
+++ b/README.md
@@ -1 +1,2 @@
 # Load, analyze, and visualize TTC bus delay data
+# Use bright colours when visualizing the data