alex21th / flight-interruption

Predictive analysis using big data technologies such as PySpark: if an unscheduled maintenance flight event is to happen sometime in the next 7 days. ✈️

Author: Jesus Antonanzas Acero, Alex Carrillo Alza

Version: "1.0"

Email: [e-mail], [e-mail]

Info: BDA, GCED, Big Data Analytics project

Date: 16/12/2019

---

Our project should contain 5 .py scripts:

• config.py configures Spark environment when necessary
• load_into_hdfs.py loads sensor CSV's into HDFS in AVRO format
• data_management.py creates training data matrix
• data_analysis.py trains a decission tree classifier model on training data
• data_classifier.py predicts new flight observations

These have to be put into the same directory as the resources directory that was included in the code skeleton, in order for the local reading option of the sensor data (CSV's) to work.

load_into_hdfs.py is to be executed first if one wants to use HDFS.

Note that the HDFS path into which the files are to be loaded has to be explicitly changed, as well as the reading path in data_management.py and data_classifier.py, if using this option.

If nothing is specified in data_management.py or data_classifier.py, though, CSV's will be read from local.

Then, the order of execution is: (optional): load_into_hdfs.py

1. data_management.py
2. data_analysis.py
3. data_classifier.py

Note that these scripts will write three objects:

1. data_matrix
2. model
3. test_matrix