Defining and setting up multimodal data in DataRobot
DataRobot's location AI, text mining, and visual AI automated ML capabilities make working with text, location, and image data relatively straightforward. With these capabilities, DataRobot can help analysts in building models and making predictions against text, image, and location-based datasets. Data set up for an image model differs considerably from other types of models. Using our House Dataset, our first task is to set up the data.
The house dataset has zip codes for the houses. We created and integrated latitude and longitude coordinates from the zip codes as new columns. Other features created from the zip codes not evident in Table 11.2 are the city, county, and state where each of the houses was located. Further text description columns were built from the number of rooms, the size of the house, and its location:
Table 11.2 – The developed price list dataset
The original data comes with 2,140 images, each of the...
