I\u2019m trying to build a proper data science\/ML project, but I\u2019m having a hard time finding a dataset that is big enough and not already used by everyone.
For example, there are datasets like the UK Online Retail dataset, Olist, and other popular sales\/retail datasets. They\u2019re good datasets, but I see them being used in a lot of projects already.
I don\u2019t want to just download a dataset, do some EDA, train a model and put it on my resume. I want to build something around an actual business problem, where I have to figure out what the problem is, analyze the data, come up with useful insights, maybe build a model, and actually explain how it could help the business.
So where do you guys usually find datasets for this?
Should I try to find data from smaller companies, government sources, APIs, research papers, etc.? Or is it okay to create my own dataset using AI\/cloud tools and then create a realistic business problem around it?
For example, if I create a large synthetic sales dataset, could I create a realistic business scenario around it and then treat it like a real project \u2014 forecasting sales, understanding customer behavior, optimizing inventory, etc.?
Would that be considered a decent portfolio project, or is using real-world data much better?
I\u2019d mainly like to hear from people who have built projects for their portfolios or have experience hiring for data science\/ML roles. Where do you actually get your data from when you want to build something that\u2019s not the same Kaggle project everyone has already done?<\/p>\n<\/div>\n
submitted by \/u\/Informal-Dog2741 <\/a>
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
