bymochimomo / fdh

Fake Data Hub is an open project for creating real looking data sets and schemas generated based on desired statistics

Fake Data Hub

Fake Data Hub is an open project for creating real looking data sets and schemas generated based on desired statistics. All code made in R!

Preface

In the modern world, we have various powerful tools to analyze data. Many of them promise a lot of advanced functionality for automatic insights, data mining, and ML. A problem comes with data: while tools do claim that they are capable to find insights in data, real-world data does not obligatory contain these insights. There are a lot of reasons for that: either we have not enough data or the data we have does not represent the population or phenomenon we are looking for, or life appear to work not the way we expected:)

To overcome this issue I decided to start a little project - Fake Data Hub. The idea is as simple as that: to get something out of the data you first need to put something in! The ultimate goal is to be able to generate any reasonable amount of fake data that looks almost like real and statistically shaped according to our needs.

Contents

1. Customers folder contains support datasets and generator code to produce people demographic and personal data. Also readydata subfolder contains generated ready-to-use data in .csv format.
2. Products folder contains support datasets and generator code to produce retail product data (partnumber, prod. categories, name, price etc). Also readydata subfolder contains generated ready-to-use product data in .csv format.
3. classification_binary folder contains genetator of binary descision Y/N to buy something based on customers data. Readydata subfolder contains ins_sales(1.5k) for 1.5k random customers and their classification based on Ins_buyers_logic.R genrator.