Hello Reader, how are you?
So glad you're well - let's dive into generating some fake data..
But first lets briefly talk about why you may want to generate fake data;
- Testing and Development: When building and testing applications, databases, or analytical models, having realistic data is essential. Fake data enables developers and testers to simulate real-world scenarios without compromising privacy or security.
- Training and Education: In educational settings, instructors and students often require datasets for learning purposes. Fake data allows educators to provide diverse, customizable datasets that facilitate hands-on learning and experimentation across various domains.
- Demonstrations and Presentations: Whether showcasing software capabilities, pitching business ideas, or conducting product demonstrations, fake data serves as a compelling tool for illustrating concepts and driving engagement.
Ok, enough of that, let's actually dive into fake data!
Starting with Mockaroo:
Let's use fake complaints as the example.
Mockaroo has a litany of different data types to explore & that likely fulfil most of your needs (& any that aren’t ChatGPT can).
Most of these data-types are fairly self-explanatory. You'll notice that you can also add Formula's (the sigma sign). This functions in the same way you'd create a formula or calculated field in most applications. A huge benefit also is just generating a large amount of rows.
Due to the eponymous nature of the options and data types the only data type I'll briefly touch on here is Customer List. This allows you to list given values that it returns randomly or based on a formula. In the above image you can see this replicated twice - to give us the Types of Complaints & the Response to Consumer.
But what is this complaint data missing?
Actual complaints.
These are difficult to make in Mockaroo, but ChatGPT is a different kind of animal.
Creating an actual complaint is rather difficult, as the complaint ideally should coincide with two other pieces of data.
The complaint needs to;
Correctly comply with the Issue types (Changed mind, other, item size issue, faulty, etc)
And the complaint must actually be a complaint about the correct product.
Fulfilling these two requirements require more input from AI to complete.
Behold below the prompts of generating those complaints:
I clearly identify for ChatGPT the data I currently have, and the fields, as well as what I'd like it to do - as a result it gives me the output I'm looking for.
After which all I need to do is copy & paste the output into the file generated by Mockaroo & I now have a complete set of fake data.