Join the 1M+ learners on CodeSignal
Be a part of our community of 1M+ users who develop and demonstrate their skills on CodeSignal
Welcome to our next course, Deep Dive into Numpy and Pandas with Housing Data. In this course, you will unlock the secrets of efficient data manipulation and analysis with Numpy and Pandas. We will build your skills from a foundational to an advanced level, strengthening your grasp of Python and preparing you for the world of Data Science.
In this first lesson, we will study the California Housing dataset. This important dataset is often used as a benchmark in machine learning and data analysis. It contains detailed information about housing values in California suburbs. The California housing market, due to its high prices and shortages, has been the subject of study for many years. This makes the dataset particularly relevant today. In this lesson, our main objective is to explore the fundamental attributes of this dataset. We aim to understand various attributes such as median income, population, average number of rooms per household, and their influence on house prices. Let's get started!
To load the California Housing dataset, we can use the sklearn library, which is powerful, easy to use, and contains many ready-to-use machine learning algorithms. It also comes with a few pre-loaded datasets, including the California Housing dataset. We can load the dataset by simply importing the appropriate sklearn module and calling a function.
After loading the California Housing dataset, we receive the data in a Bunch object; it's similar to a dictionary but with added functionalities. It has keys like data, target, feature_names, each leading to a different part of the dataset. The data key contains all the input features, the target key has the output values we might want to predict (median house values), and feature_names holds the names of the features.
A pandas DataFrame is more familiar and offers more functionality, making it easier to work with. Let's convert our dataset into a pandas DataFrame.
This code converts the different parts of the Bunch object into a Pandas DataFrame. We also add the target values to the dataframe as a new column, MedHouseValue. Now, we have successfully imported our dataset and already explored the first few rows using the head() function!
Once we've imported the dataset, it's time for some preliminary analysis to get a general understanding of our data.
The shape attribute gives us the dimensions of our dataset - the number of rows and columns. The describe() function provides valuable statistical information about each attribute, including the count, mean, standard deviation, and various percentiles. With this information, we can understand the distribution of data across each attribute.
To make better use of our dataset, we need to understand the features it includes. Every feature or column in our dataset represents a characteristic of a block group in California. These characteristics include:
MedInc: This is the median income for households within a block (scaled and capped at 15 for higher median incomes and at 0.5 for lower median incomes).HouseAge: This is the median house age within a block.AveRooms: This is the average number of rooms in the houses within a block.AveBedrms: This is the average number of bedrooms in the houses within a block.Population: This is the total population within a block.AveOccup: This is the average house occupancy, computed as the total population within a block divided by the number of households.Latitude and Longitude: These are the geographic coordinates of the block groups.MedHouseValue: This is the median house value for households within a block (measured in 100,000s).Understanding these variables allows us to make more sense of our dataset and prepare it for machine-learning tasks.
Knowing your data is crucial, but it is equally important to ensure our data is clean and ready for machine learning models. Data preprocessing involves a wide range of tasks, such as dealing with missing values, data normalization or standardization, handling outliers, etc.
Before we preprocess, we need to check if we have any missing values. With pandas, it's as easy as calling the isnull() function.
The above line of code checks for missing values in our dataset and sums up the null values found for each attribute. As we can see, our dataframe doesn't contain any missing values.
This lesson introduced you to the process of importing a dataset from the sklearn library, understanding its basic characteristics, investigating each feature, and performing a visual inspection of the data. You've also learned why data cleanup is important before applying machine learning models. Our journey with the California Housing dataset has just begun, and it's already proved to be a vital experience for understanding and working with real-life data.
Well done on completing this lesson! Now, it's time to reinforce your understanding and familiarize yourself even more with the California Housing data. Up next are practice exercises designed to help you internalize data handling and manipulation aspects of Python, Numpy, and Pandas.