From the course: Artificial Intelligence Foundations: Machine Learning

Obtaining data

From the course: Artificial Intelligence Foundations: Machine Learning

Obtaining data

“

Data is a critical element of any machine learning project. There are many places from where you can source data. You may have an internal data store you can query, your client may provide the data, an open source or public data store could help, or you may need to buy data from a third party. Do you ever dream about buying a home in California near the ocean? I know I do. If you do, you'll want to study these cost predictions. We'll use information regarding the demography like income, population, house occupancy in the districts, the location of the districts, latitude and longitude, and features of the home like number of bedrooms, number of rooms, age of the house. We'll use that to predict what that home will sell for. We'll use a machine learning technique called supervised learning to predict the cost of the home, which is a linear regression problem. Do you remember the definition of supervised learning from Chapter 1? If not, let's refresh. Supervised learning is a type of machine learning where machines are trained using labeled data. Labeled data is data that already contains the target value the machine needs to learn how to predict. In this case, the target is the median house value which is present in the data. Therefore, the machine reviews the other features, also called variables, and learns how to predict the median house value, also called the target. Linear regression is a machine learning algorithm based on supervised learning. Regression models often predict a numeric value called a target, in our case, the cost of the home, based on understanding the relationships between independent variables, the features. Linear regression is mostly used to find the relationship between variables and forecasting. Let's take a closer look at the housing dataset. There are several feature variables. First, there's longitude. Longitude is a measure of how far west a house is. A higher value is farther west. Next, we have latitude in this column. It's a measure of how far north a house is. A higher value is farther north. Next we have the housing median age column. This is the median age of a house within a block. A lower number is a newer building. Next we have total rooms. Total rooms is the total number of rooms within a block. The next column is total bedrooms. It's the total number of bedrooms within a block. We have population. It is the total number of people residing within a block. And let's scroll over just a little more. In the G column, we have households. And that is the total number of households, essentially a group of people, residing within a home unit for a block. Median income, I'm sure you can guess what that is. It's the median income for households within a block of houses, and it's measured in tens of thousands of US dollars. I'm going to skip over median house value, and let's talk about ocean proximity. This is the location of the house in proximity to the ocean. Now I'm going to go back to median house value because it is the most important feature in this dataset. First, the median house value is what that home is worth. Why is this field so important? It's important because it's the target variable. This is the field that we want the machine to learn how to predict. Now let's review the Python code. The first step is to load the data set, and that's what we're doing here using this read CSV function. I'm loading the CSV file data into a pandas data frame. After loading the data, I can now access it through this housing DF, which is the data frame. Now that we've loaded the data set, we're ready to explore it by generating 2D graphs using Matplotlib and Seaborn.

Contents