Consider data which has following columns:
1. Price of land.
2. Distance from hospital
3. Distance from fire station
4. Traffic on roads
5. Number of red-colored 2-wheeler.
Here intuitively, the cost of land will vary as per distance from the hospital, distance from the fire station, traffic on roads,
however, it shall not depend much on number of red-colored 2-wheeler.
The English word “model” is a beautiful thing in that sense.
If you look at the dictionary meaning of model, “a copy of something that is usually smaller than the real thing”.
So the model depicts the patterns/relationships already inherent in data.
In that sense Model/Algorithm, is not something to which when data is passed, magical things start coming up.
In fact, calling it an algorithm is also a wrong an analogy because algorithm is something that solves a problem.
Model is the right word for it.
All mystery is in the data. If the data has certain relationships/correlations, the model simply learns those and the next time you ask it a question it gives you an answer on similar lines.
So, for aspiring Data Scientists, it’s important to learn and understand algorithms but diving deep into data is a must.
Understanding relationships/correlations in the data is foremost essential, and once you understand that you shall easily be able to work on appropriate models.
codeSimplfied 