So I’ve got a nutty little project that I think only Hackeet is capable of helping me with. I’m posting this as another learning journey. Even if the project is a dead end, it’ll be a really interesting experiment.
So I’ve got a side-project with a colleague in the restaurant business. She’s interested in trying to take reservation data (first + last name, email, phone) to see if there’s anything public available on the internet about the guest, such as their profile photo to more easily identify them when they arrive at the restaurant, or their Instagram profile to see if they’ve posted about something they like to drink. Basically a ‘get to know your guest’ tool for customers who are active publicly on the internet.
This is a big challenge in terms of APIs. We’re in the age of increasing privacy, so entering someone’s email into Google doesn’t cut it anymore. We need to be smart if we want to discover public information about someone without invading their privacy. There are a few ways this can be achieved:
-
What’s the domain name of their email address?
1.a. If it’s gmail, use Google People API to see if they have a profile pic
1.b. If it’s not gmail or yahoo or hotmail, is it a custom domain that points to their place of work or study?
1.c. If the domain extension isn’t .com but is like .co.uk or .fr that would help us narrow down results later by country. -
What’s the first part of their email address?
2.a. It could be a username, like richardosborne14 for me which I often use as a plain username for websites that require it. -
Ask the customer to post about their reservation on Instagram or Twitter with a specific hashtag (like a share button with a pre-written share text including the hashtag)
3.a. Hashtags are searchable on Instagram and Twitter APIs, so we can find their public social profile and pull their public posts for useful info -
Phone - not sure how to best use this yet.
-
First and last name - search on Google Custom Search Engine and see if they’re a public personality. Perhaps they have a popular blog, or their social profile is popular enough to be highly ranked.
5.a. This would only work with either human intervention to check results, or a strong AI analysis like ChatGPT
5.b. Check the result for any information that can help us identify the customer
That’s what I’ve got as ideas so far, and there will likely be more. I’ve tried to map out the ideas in Miro to show how I think Hackeet will achieve working with all these different APIs, receiving and parsing the data, sending it to ChatGPT and dealing with the analysis results, etc.
To be continued!