At our last meeting, we discussed removing redundant copies of providers. It improves data integrity, allow us to delete about 100 rows, which means less data to cache on the client side, and less bumping up against the 1000-row limit.
Sometimes it's not clear whether two rows should be collapsed together, for example, if everything is the same except the description, and the descriptions are customized for the category/subcategory, it should probably be kept as two separate items.
Sometimes it's less clear. For example, G-Street Oasis has 8 entries, where one (laundry) had a different phone number than the others. But other than that all 8 entries were identical. We kept the laundry one as a separate row, and collapsed the other 7. I would want to double check that - maybe the laundry one just had the wrong phone number?
To handle situations like this, I suggest adding comment column to providers table for use in data cleaning. In the above situation, I would add a comment saying something like: "Is there really a separate phone number for laundry, or should this be collapsed with G-Street Oasis?"
Then, as a separate function, we can have a person periodically search through these comments and resolve them.
Here's just another example, of which there are many: There are "ACLU" and "American Civil Liberties" items that share an address and phone number (but one has an additional number), but they have different descriptions. I would use the comment column and say something like "I'm thinking pick a name for both, and give them both the same phone numbers, and they have different lines because they have different descriptions."
At our last meeting, we discussed removing redundant copies of providers. It improves data integrity, allow us to delete about 100 rows, which means less data to cache on the client side, and less bumping up against the 1000-row limit.
Sometimes it's not clear whether two rows should be collapsed together, for example, if everything is the same except the description, and the descriptions are customized for the category/subcategory, it should probably be kept as two separate items.
Sometimes it's less clear. For example, G-Street Oasis has 8 entries, where one (laundry) had a different phone number than the others. But other than that all 8 entries were identical. We kept the laundry one as a separate row, and collapsed the other 7. I would want to double check that - maybe the laundry one just had the wrong phone number?
To handle situations like this, I suggest adding comment column to providers table for use in data cleaning. In the above situation, I would add a comment saying something like: "Is there really a separate phone number for laundry, or should this be collapsed with G-Street Oasis?"
Then, as a separate function, we can have a person periodically search through these comments and resolve them.
Here's just another example, of which there are many: There are "ACLU" and "American Civil Liberties" items that share an address and phone number (but one has an additional number), but they have different descriptions. I would use the comment column and say something like "I'm thinking pick a name for both, and give them both the same phone numbers, and they have different lines because they have different descriptions."