[Bug Report] xgboost_customer_churn.ipynb fails

**Link to the notebook**
https://github.com/aws/amazon-sagemaker-examples/blob/master/introduction_to_applying_machine_learning/xgboost_customer_churn/xgboost_customer_churn.ipynb

**Describe the bug**
Running Cell [7]:
```
# Frequency tables for each categorical feature
for column in churn.select_dtypes(include=["object"]).columns:
display(pd.crosstab(index=churn[column], columns="% observations", normalize="columns"))

# Histograms for each numeric features
display(churn.describe())
%matplotlib inline
hist = churn.hist(bins=30, sharey=True, figsize=(10, 10))
```
The result is the following error:
```
TypeError: Cannot interpret '' as a data type
```

RCA: [pandas <1.0.5 is not compatible with numpy 1.20](https://github.com/hdmf-dev/hdmf/issues/617)
On SageMaker Studio - datascience-1-0-ml-t3-medium:
`pandas==1.0.1`
`numpy==1.20.3`

**To reproduce**
Run the notebook.

**Logs**
```
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
in
1 # Frequency tables for each categorical feature
2 for column in churn.select_dtypes(include=["object"]).columns:
----> 3 display(pd.crosstab(index=churn[column], columns="% observations", normalize="columns"))
4
5 # Histograms for each numeric features

/opt/conda/lib/python3.7/site-packages/pandas/core/reshape/pivot.py in crosstab(index, columns, values, rownames, colnames, aggfunc, margins, margins_name, dropna, normalize)
579 from pandas import DataFrame
580
--> 581 df = DataFrame(data, index=common_idx)
582 if values is None:
583 df["__dummy__"] = 0

/opt/conda/lib/python3.7/site-packages/pandas/core/frame.py in __init__(self, data, index, columns, dtype, copy)
433 )
434 elif isinstance(data, dict):
--> 435 mgr = init_dict(data, index, columns, dtype=dtype)
436 elif isinstance(data, ma.MaskedArray):
437 import numpy.ma.mrecords as mrecords

/opt/conda/lib/python3.7/site-packages/pandas/core/internals/construction.py in init_dict(data, index, columns, dtype)
252 arr if not is_datetime64tz_dtype(arr) else arr.copy() for arr in arrays
253 ]
--> 254 return arrays_to_mgr(arrays, data_names, index, columns, dtype=dtype)
255
256

/opt/conda/lib/python3.7/site-packages/pandas/core/internals/construction.py in arrays_to_mgr(arrays, arr_names, index, columns, dtype)
67
68 # don't force copy because getting jammed in an ndarray anyway
---> 69 arrays = _homogenize(arrays, index, dtype)
70
71 # from BlockManager perspective

/opt/conda/lib/python3.7/site-packages/pandas/core/internals/construction.py in _homogenize(data, index, dtype)
321 val = lib.fast_multiget(val, oindex.values, default=np.nan)
322 val = sanitize_array(
--> 323 val, index, dtype=dtype, copy=False, raise_cast_failure=False
324 )
325

/opt/conda/lib/python3.7/site-packages/pandas/core/construction.py in sanitize_array(data, index, dtype, copy, raise_cast_failure)
463 value = maybe_cast_to_datetime(value, dtype)
464
--> 465 subarr = construct_1d_arraylike_from_scalar(value, len(index), dtype)
466
467 else:

/opt/conda/lib/python3.7/site-packages/pandas/core/dtypes/cast.py in construct_1d_arraylike_from_scalar(value, length, dtype)
1459 value = ensure_str(value)
1460
-> 1461 subarr = np.empty(length, dtype=dtype)
1462 subarr.fill(value)
1463

TypeError: Cannot interpret '' as a data type
```

Open introduction_to_applying_machine_learning/xgboost_customer_churn/xgboost_customer_churn.ipynb and run Cell [7] in the reported SageMaker Studio environment. Check the pandas==1.0.1 and numpy==1.20.3 compatibility described in the report; done means the cell runs without the reported TypeError and still displays the frequency tables, summary, and histograms.

Written by the indexing model from the issue text.