Machine Learning
How complicated is it to use data science?
Using data science becomes easier every year. Great software tools like Scikit-Learn and Google’s TensorFlow combined with Keras make it possible to code very advanced models with just 50 lines of code, often even less. These tools are free and can be used by individuals and companies.
Who do we identify as our primary data science audience?
Finance and controlling:
Transaction data in organizations is vast. Controllers and consultants need to scan and analyze that data with advanced machine learning methods today in order to gain new interesting insights. Otherwise a great deal of the effort to collect and store data is wasted. For example, fraud detection or insurance risk classification can be performed much better by machine learning algorithms than by humans.
More detailed examples for banks and insurances
Engineering:
Sensor data is everywhere nowadays. In order for it to be useful it has to be analyzed and ideally action is taken in real time. For example, machines take action autonomously on the basis of real time image analysis. Or predictive maintenance: Machines are being observed by sensors and on the basis of that data good next time inspections are derived and planned.
More detailed engineering example applications
Marketing:
Identifying the right customers to target is essential for a successful marketing campaign. By analyzing customer behavior data using machine learning algorithms this can be done.
Another example is churn prevention: Identifying those customers which are likely to leave. Identifying such customers on the basis of customer behavior data in time, action can be taken to keep these customers.
More detailed example applications for marketing
Medicine:
Computer diagnosis systems nowadays work on the basis of data of past patients and image recognition programs. All such systems use machine learning to a great extent.