First Normal Form (1NF) states that a single field should store a single data point and that all rows should be unique.

The important point here is that your data should be atomic, so large tables with repeated data should be broken out into smaller tables with easily searchable keys. This makes filtering more efficient and querying more straightforward. It literally makes your job is easier!,This tutorial explains how to preprocess data using the Pandas library. Preprocessing is the process of doing a pre-analysis of data, in order to transform them into a standard and normalised format. Preprocessing involves the following aspects:,In the remainder of the tutorial, we apply each method to a single column. However, if you wanted to use each column of the dataset as input features of a machine learning algorithm, you should apply the same normalisation method to all the columns.,Third Normal Form (3NF) states that no functional transitive dependencies are allowed. This one might be the most complicated, but essentially it means that if you have a field that is dependent on another field, it should be broken out to another table. This allows the foreign key to change — not the actual value.,Second Normal Form (2NF) states that each table should have a single primary key column. Create unique IDs. Instead of using name and address as a compound key, just give the user a unique ID number or GUID.,Data Normalisation involves adjusting values measured on different scales to a common scale. When dealing with dataframes, data normalization permits to adjust values referred to different columns to a common scale. This operation is strongly recommended when the columns of a dataframe are considered as input features of a machine learning algorithm, because it permits to give all the features the same weight.,First of all, we need to import the Python pandas library and read the dataset through the read_csv() function. Then we can drop all the columns with NaN values. This is done through dropna() function.,I’m biased because this has worked well for me, but that doesn’t mean that I can’t recommend blogging! With a platform like Medium, you can write project walkthroughs, like mine on Wine Quality Prediction.,I hope that this provides some direction and helps you in your data science career. There’s no cookie-cutter way of approaching this, so feel free to take this with a grain of salt. Nonetheless, I wish you the best in your data science endeavors!,In this tutorial, we use the pandas library to perform normalization. As an alternative, you could use the preprocessing methods of the scikit-learn libray. A little note for readers: if you wanted to learn how to use the preprocessing package of scikit-learn, please drop me a message or a comment to this post :),Lastly, take advantage of non-profit data science opportunities. I came across a resourceful article written by Susan Currie Sivek, which provides several organizations where you can get the opportunity to work on real-life data science projects.,As example dataset, in this tutorial we consider the dataset provided by the Italian Protezione Civile, related to the number of COVID-19 cases registered since the beginning of the COVID-19 pandemic. The dataset is updated daily and can be downloaded from this link.,Synopsis: “Netflix’s new series ‘The Baby-Sitters Club’ is based on the best-selling book series, that follows the friendship and adventures of Kristy Thomas, Mary-Anne Spier, Claudia Kishi, Stacey McGill, and Dawn Schafer as the middle-schoolers start their babysitting business in the town of Stoneybrook, Connecticut.”,If you are in a more full-stack role or with a company that doesn’t have a database team, you will likely be responsible for database development. In this case, you should have a good idea of which kinds of database objects to use in different scenarios.,I hope that this provides some direction and helps you in your data science career. There’s no cookie-cutter way of approaching this, so feel free to take this with a grain of salt. Nonetheless, I wish you the best in your data science endeavors!