Step 1: We will gather information (in Excel format) on every vehicle sold on Carwale.com.
Step 2: Following data collection, we will clean the data, which entails either deleting any missing or corrupted values or attempting to locate them every month or using statistics.
Step 3: Feature engineering is what we’ll do next. We will exclude odd fields that have no bearing on the selling price because we are aware of the columns in our data, which include the car’s name, mileage driven, number of owners, power, and, last but not least, the selling price; reduce the data; convert alphabetic columns (vehicle names) to integer form because machines can only comprehend numbers; Plot column graphs to observe.

Step 4: Divide the data into training and testing sets (much to how our maths teacher sets aside some problems for the homework test and solves some number problems in class).
Step 5: This is when ML comes into play. A machine learning method will be chosen so that the computer can recognise patterns in the data. (Don’t worry about how to choose; we’ll talk about how the algorithm looks for patterns later.) We employ diverse approaches to master different things; for example, we don’t use the same approach for maths and social studies. Machines learn data using various algorithms, just like humans do.

Step 6: Following instruction, testing will be conducted. We will provide the machine with automobile details during testing, and it will estimate the cost.

Step 7: Lastly, we’ll store the knowledge. (Again, ignore the type of learning; just suppose that, like computers, our brains store information in complicated form.) We shall talk about that later), and buyers may use that information to forecast the price of their cars on the website.
Every action listed above is data science.

The steps will change based on the assignment. The procedures will alter if image data is available, but the general task remains the same.

Crucial point: Machine learning is limited to prediction. There are many of data science projects that don’t use machine learning. ML is thus merely a tool that can be employed or not.
Returning to data science in mechanical engineering, let’s talk about it now.
Approximately 90% of jobs in the manufacturing, metallurgy, oil and gas, automotive, biomedical, and defence sectors can be automated with data science, from designing to selling. Data science can generate appropriate designs, monitor factory performance, and anticipate machine failures before they cause downtime. For this reason, businesses are cutting employees in order to lower production costs. Recessions occur for this reason.

Some news stories and articles from Forbes and the companies’ original websites bolster these claims: