Preface
1
Introduction
1.1
Why
1.2
What
1.3
How
2
The Data
3
Visualizations
4
Forecasting Models
4.1
PM10 forecasts
ARIMA results
Machine Learning results
4.2
NO2 forecasts
5
Conclusions
R scripts
6
Gathering and Cleaning Data
6.1
Data gathering
6.2
Data cleaning
6.3
Adding new variables
6.3.1
Time variables
6.3.2
Laboral dates
6.3.3
Wind direction
6.3.4
Tables preparation for Tableau dashboards
7
Data Exploration
7.1
Trends exploration
7.2
PM10 Constitucion Station
7.3
NO2 Constitucion Station
7.4
Relationships between variables
8
Forecasting models. ARIMA
8.1
Data loading
8.2
Train, test and validation data (PM10 models)
8.3
Data Exploration
8.4
Base model.
8.5
PM10 ARIMA models
8.5.1
PM10_model_arima_3m
8.5.2
PM10_model_arima_3y
8.5.3
PM10_model_arima_9y
8.5.4
PM10 h = 6
8.5.5
PM10 h = 12
8.5.6
PM10 h = 24
Python scripts
Gijón Air Pollution - An exercise of visualization and forecasting
R scripts
I appended in the next three chapters the following R scripts:
10_Gathering_and_Cleaning_Data.rmd
11_Data_Exploration.rmd
12_Forecasting_Models_ARIMA_PM10.rmd