OUR PHILOSOPHY
To make informed business decisions, start by articulating a clear objective, such as predicting customer churn or optimizing pricing. Identify the essential data needed for the decision-making process, encompassing relevant variables and their sources. Appreciate the nature of the variables instead of just their numbers to improve dramatically performance. Recognize that any time spent coding is a cost, while all time spent on experimentation and analysis yield valuable benefits. Embrace the Dual-Experiment Data Mining Process Flow to achieve optimal accuracy with minimal effort. Cultivate expertise in modeling economics over algorithmic intricacies by efficiently focusing on the business problem, adhering to the philosophy that judiciously allocated resources in experimentation and analysis contribute significantly to achieving high-performance data-driven decision making.
Start with the business decision and work backward – what do you need to know and what is nice to know?Â
When preparing predictor for modeling, it’s all or nothing                             Â
All time spent coding is a cost and any time spend experimenting & analyzing is benefit.
Invoke the Dual Experiment Data Mining Process Flow to get the highest accuracy with the least work
Gain expertise in the modeling economics instead of algorithms.
OUR PROMISE
This book contains our citizen data science philosophy, theories of machine learning explained in plain English, practical applications for many types of predictive and interpretable models, and the software and datasets required for doing real data science without becoming a data scientist.
The data mining process flows, hosted in Jupyter notebooks along with the datasets in the cloud, are complete, and templates are fully populated for all main types of models. In fact, this book is where template-driven data analytics gets its start, but you will decide where it ends as it can be applied to limitless situations. While the backend technology is very advanced, you can use a Chromebook to build, run, and analyze professional-grade machine learning models using this guide.
Providing predictive and interpretable models with detailed instructions on analyses is the first step. Future editions of this book are anticipated that use no-code PyCaret for time series forecasting and unsupervised models, such as clustering.
Individuals can benefit from this book by becoming more valuable in a wide variety of settings, but so can organizations – if you work in a medium-to-large company, please see the next section on the organizational reasons for accepting citizen data science as part of company-wide analytics.
