After running count(distinct(checkin_dates)), the output found is 348, which is weird because the date range in check-ins is from January 2010 to December 2019. Suspect that some days went missing in one of the tables. This issue also caused a gap between # of check-in per day in this chart:

Unequal number of rows in raw_review and review table

--Creating review table
CREATE TABLE review (review_id string, rev_user_id string, rev_business_id string, rev_stars int, rev_useful int, rev_funny int, rev_cool int, rev_text string, rev_date string);
--Populating review table from raw_review
FROM raw_review INSERT OVERWRITE TABLE review SELECT get_json_object(json_response, '$.review_id'), get_json_object(json_response, '$.user_id'), get_json_object(json_response, '$.business_id'), get_json_object(json_response, '$.stars'), get_json_object(json_response, '$.useful'), get_json_object(json_response, '$.funny'), get_json_object(json_response, '$.cool'), get_json_object(json_response, '$.text'), get_json_object(json_response, '$.date');

Some of the reviews have quotation marks and escape characters. These reviews are not being parsed correctly. For example, one review's text attributes look like:

"As someone who has worked with many museums, I was eager to visit this gallery on my most recent trip to Las Vegas. When I saw they would be showing infamous eggs of the House of Faberge from the Virginia Museum of Fine Arts (VMFA), I knew I had to go!\n\nTucked away near the gelateria and the garden, the Gallery is pretty much hidden from view. It's what real estate agents would call "cozy" or "charming" - basically any euphemism for small.\n\nThat being said, you can still see wonderful art at a gallery of any size, so why the two s you ask? Let me tell you:\n\n pricing for this, while relatively inexpensive for a Las Vegas attraction, is completely over the top. For the space and the amount of art you can fit in there, it is a bit much.\n* it's not kid friendly at all. Seriously, don't bring them.\n* the security is not trained properly for the show. When the curating and design teams collaborate for exhibitions, there is a definite flow. That means visitors should view the art in a certain sequence, whether it be by historical period or cultural significance (this is how audio guides are usually developed). When I arrived in the gallery I could not tell where to start, and security was certainly not helpful. I was told to "just look around" and "do whatever." \n\nAt such a fine institution, I find the lack of knowledge and respect for the art appalling."

After parsing, the review table's rev_text column has only this portion:
As someone who has worked with many museums, I was eager to visit this gallery on my most recent trip to Las Vegas. When I saw they would be showing infamous eggs of the House of Faberge from the Virginia Museum of Fine Arts (VMFA), I knew I had to go!

And, the rev_date column became null.

It took 11 rows to completely parse one review. Similar things happened for all the reviews with \n with it.

As a result, 8,021,122 rows from the raw_review table became 25,129,230 rows in the review table.

We can fix this issue by somehow ignoring '\n' when populating the review table.