Data science is a collection of Data inference, algorithm development and technology, focusing on to solve analytically complex problems. At the center, there is data, there are more valuable raw information streaming through the Internet and stored in enterprise data warehouses, what data science doing is, using all that data in more intelligent and creative ways to process and produce insights/information which is used to generate business value, in other way data science helps us to do quantitative data analysis, to make strategic business decisions, just about anything.
There is a side of data science, it's all about uncovering findings from data, it's called "discovery of data insight". whats happening with this is, we look at data in the smallest level we can, and with all the data mining tools, we can understand complex behaviors, trends. It's about revealing hidden information that can help users and companies(mostly companies) to make smarter business decisions. Best example is "Netflix", Netflix is a media streaming company which produce and streams movies and television series, with data science, they are uncovering data of movie viewing patterns to understand what users interested in, and using that information to make decisions on "which Netflix original series" to produce.
It happens with "data exploration" done by data scientists. When given a challenging and complex question, those data scientists become more like "detectives". They investigate leads and understand patterns or characteristics in the data, this requires a huge amount of analytical creativity. After that if needed, data scientists may apply quantitative techniques to get more deep in to the problem, such as "time series forecasting", "segmentation analysis", "synthetic control experiments". The main priority is to scientifically bring together a forensic view of what the data is really saying. This data-driven information/insights or the outputs, are the base, to provide strategic guidance. like detectives, data scientists act as consultants providing guidance to business stakeholders on how to act on data findings.
Another side of the data science is the "development of data product", as a term a "data product" is a technical asset that utilizes data as input, and processes that data to return algorithmically-generated results, like recommendation engines. "recommendation engine" is a data product which absorb user data and makes personalized recommendations based on them. Google's "Gmail spam filter" is a data product, in behind, a clever algorithm processes incoming mail and determines if that message is a spam or not.
But the "data product" is different from the "data insight" in data insights, the outcome of it provide advice to a business stakeholder to make a business decision, on the other hand a data product is a functionality includes algorithms, and designed to integrate directly in to core applications like "spam filter on Gmail's inbox", "Tesla's autonomous driving software". On developing data products, data scientists playing a key role in building out algorithms, testing and making refinements, they also involved in technical deployment of data products, into production systems. In this part the data scientists are serve as technical developers, designing and building assets that can be used in more wider scope.