Hands-On Deep Learning for IoT

Exploratory analysis

To give an idea of the sensor readings in areas such as the physical state of the engine (for example, with regard to the temperature of a component, the fan speed of the turbine, and so on) we decided to extract the first unit from the first dataset for all the sensors on a single engine. For this, we have written a script (see make_dataset.py) that gets all of the data files from the input directory. Then it parses a set of raw data files into a single DataFrame object and returns an aggregated representation of all files with the appropriate column names:

data_sets = []    for data_file in glob(file_pattern):        if label_data:            # read in contents as a DataFrame            subset_df = pd.read_csv(data_file, header=None) # need to create ...