The fundamentals for understanding what Data Science is [Definition]
Posted: Sat Jan 04, 2025 7:18 am
The word “Data Science” has become quite common. But what is it exactly? What skills are required to become a Data Scientist? What is the difference between Business Intelligence and Data Science? How are decisions and predictions made in Data Science? Here are some of the questions we will try to answer.
Let's start with the basics. What is the definition of "Data Science"? We already touched on it a few words in the introduction. Very schematically, we could say that Data Science mixes tools, algorithms and machine learning principles with the aim of discovering hidden patterns from raw data. But how is the work of a Data Scientist different from what data analysts have been doing for years? The whole difference between the two comes down to the difference between explaining and predicting.
definition of data science data analysis vs data science
The Data Analyst explains what happens when exploring data. To discover insights, the Data Scientist, on the other hand, does not just use exploratory analysis, he also uses advanced machine learning algorithms to identify occurrences of this or that event in the future. A Data Scientist looks at data from several angles, some of which are previously unknown. Data Science is mainly used to make decisions and predictions based on causal predictive analysis, normative analysis and machine learning. Let's take a look at each of these terms:
Predictive causal analytics . If you want a model that can predict the probability of an event occurring in the future, you need to do predictive causal analytics. If your business is lending money, knowing whether your paraguay whatsapp list customers will pay you back on time is something that you are concerned about. If this is your case, you could build a model to perform predictive analytics and predict whether your customers will pay you back on time, based on their payment history. Predictive analytics is an important dimension of Data Science.
Prescriptive analytics . If you want a model that has the intelligence to make decisions on its own and the ability to modify them based on dynamic parameters, you will definitely need what is called prescriptive analytics. This field is relatively new. The goal is to generate recommendations. In other words, the model does not just predict but suggests several possible actions with their predictable results. The best example of the application of prescriptive analytics is self-driving cars. The data collected by the vehicles can be used to “train” self-driving cars. You can run algorithms on this data to bring intelligence to it. The result of the analyses is then used to make decisions: turn, accelerate, slow down, etc.
Machine learning (for decision making) . If you have a company’s transactional data in finance at your disposal and you need to build a model to determine future trends, then machine learning is probably the best solution. In this case, it is called supervised learning. “Supervised” in the sense that you already have the data in advance that will be used to train the machines. For example, you can train a fraud detection model using the history of fraudulent purchases.
Machine learning (to discover patterns). If you do not have the necessary parameters in advance to make the predictions (as in the previous case), you must find hidden patterns (= motifs) by exploring the database. This is called unsupervised models. Learning is done "freewheeling", without human guidance. Most of the time, a clustering algorithm is used to discover hidden patterns. Let's take an example. You are a telephone company. You need to create a network of pylons in a region. You can use the clustering technique to identify the best placement of the pylons (= the one allowing network users to obtain the best signal). This is one example among many others... On the difference between supervised learning and unsupervised learning, we invite you to read this article , very clear and very interesting.
Let's start with the basics. What is the definition of "Data Science"? We already touched on it a few words in the introduction. Very schematically, we could say that Data Science mixes tools, algorithms and machine learning principles with the aim of discovering hidden patterns from raw data. But how is the work of a Data Scientist different from what data analysts have been doing for years? The whole difference between the two comes down to the difference between explaining and predicting.
definition of data science data analysis vs data science
The Data Analyst explains what happens when exploring data. To discover insights, the Data Scientist, on the other hand, does not just use exploratory analysis, he also uses advanced machine learning algorithms to identify occurrences of this or that event in the future. A Data Scientist looks at data from several angles, some of which are previously unknown. Data Science is mainly used to make decisions and predictions based on causal predictive analysis, normative analysis and machine learning. Let's take a look at each of these terms:
Predictive causal analytics . If you want a model that can predict the probability of an event occurring in the future, you need to do predictive causal analytics. If your business is lending money, knowing whether your paraguay whatsapp list customers will pay you back on time is something that you are concerned about. If this is your case, you could build a model to perform predictive analytics and predict whether your customers will pay you back on time, based on their payment history. Predictive analytics is an important dimension of Data Science.
Prescriptive analytics . If you want a model that has the intelligence to make decisions on its own and the ability to modify them based on dynamic parameters, you will definitely need what is called prescriptive analytics. This field is relatively new. The goal is to generate recommendations. In other words, the model does not just predict but suggests several possible actions with their predictable results. The best example of the application of prescriptive analytics is self-driving cars. The data collected by the vehicles can be used to “train” self-driving cars. You can run algorithms on this data to bring intelligence to it. The result of the analyses is then used to make decisions: turn, accelerate, slow down, etc.
Machine learning (for decision making) . If you have a company’s transactional data in finance at your disposal and you need to build a model to determine future trends, then machine learning is probably the best solution. In this case, it is called supervised learning. “Supervised” in the sense that you already have the data in advance that will be used to train the machines. For example, you can train a fraud detection model using the history of fraudulent purchases.
Machine learning (to discover patterns). If you do not have the necessary parameters in advance to make the predictions (as in the previous case), you must find hidden patterns (= motifs) by exploring the database. This is called unsupervised models. Learning is done "freewheeling", without human guidance. Most of the time, a clustering algorithm is used to discover hidden patterns. Let's take an example. You are a telephone company. You need to create a network of pylons in a region. You can use the clustering technique to identify the best placement of the pylons (= the one allowing network users to obtain the best signal). This is one example among many others... On the difference between supervised learning and unsupervised learning, we invite you to read this article , very clear and very interesting.