ColumnSchema utils `is_numeric` and `is_categorical` only work if logical type is set

Currently the ColumnSchema utils `is_numeric` and `is_categorical` only look at the logical type when determining if a column schema is numeric or categorical in nature. In Featuretools, we often set just a `"numeric"` or `"category"` semantic tag in a column schema with no logical type in order to have the column schema be more general. Currently, those situations return `False` when checking `is_numeric` and `is_categorical` respectively.

Ideally, Woodwork would recognize both of those as numeric and categorical as follows:

```python
col_schema = ColumnSchema(semantic_tags={'numeric'})
assert col_schema.is_numeric

col_schema = ColumnSchema(semantic_tags={'category'})
assert col_schema.is_categorical

```