Improve performance of reading data from CSV files

Once #1307 is complete, we should investigate different approaches for speeding up the process of deserializing data from CSV files.

This could possibly be used for deserialization with the `read_woodwork_table` function as well as in the `read_file` function.

Some resources that may be helpful:
- https://pythonspeed.com/articles/pandas-read-csv-fast/
- https://ursalabs.org/blog/fast-pandas-loadi
- https://towardsdatascience.com/apache-arrow-read-dataframe-with-zero-memory-69634092b1a