`pivot_table()` to `groupby().agg()`

**Instance of the original version**
```python
# Example:
df.pivot_table(index='Survived', values='Age', aggfunc='sum')
```

**Instance of the faster version**
```python
# Example:
df.groupby('Survived').agg({'Age': 'sum'})
```

This is about 3x faster on the titanic dataset. We need to test more.
