Refactor: centralize column constants and eliminate duplicate segment assignment

[Copilot]

Two classes of code duplication existed across the src/ modules: column name strings were hardcoded independently in both feature_engineering.py and data_preprocessing.py, and df['segment'] = labels was assigned in both generate_model.py and profiling.py.

Changes

• New src/constants.py — single source of truth for all column names (NUMERICAL_COLS, CATEGORICAL_COLS, and individual constants like ENGAGEMENT_SCORE_COL, CTR_COL, etc.)

• src/data_preprocessing.py — replaces inline column lists with imports from constants.py

• src/feature_engineering.py — replaces hardcoded column string literals with named constants from constants.py

• generate_model.py — removes duplicate df['segment'] = labels assignment; delegates to profile_segments(df, labels) which already performs it

# Before — segment assigned in two places
labels = cluster_users(X, k=4)
df['segment'] = labels          # ← duplicated; also done inside profile_segments()

# After — single assignment via existing function
labels = cluster_users(X, k=4)
profile_segments(df, labels)    # owns the df['segment'] assignment

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.