Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Calidad del aiRe de Madrid (2001-2016)

Actualización de "Pongamos que hablo del aiRe de Madrid".
by

Carlos Ortega

on 14 January 2016

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Calidad del aiRe de Madrid (2001-2016)

Calidad del ai
R
e de Madrid
2001-2015

The Theil-Sen method dates back to 1950, but the basic idea pre-dates 1950 (Theil 1950; Sen 1968). It is one of those methods that required the invention of fast computers to be practical.
The basic idea is as follows. Given a set of n x, y pairs, the slopes between all pairs of points are calculated. Note, the number of slopes can increase by ≈ n 2 so that the number of slopes can increase rapidly as the length of the data set increases.
The Theil-Sen estimate of the slope is the median of all these slopes.
The advantage of the using the Theil-Sen estimator is that it tends to yield accurate confidence intervals even with non-normal data and heteroscedasticity (non-constant error variance). It is also resistant to outliers — both characteristics can be important in air pollution.
The smoothTrend function calculates smooth trends in the monthly mean concentrations of pollutants.
In its basic use it will generate a plot of monthly concentrations and fit a smooth line to the data and show the 95 % confidence intervals of the fit.
The smooth line is essentially determined using Generalized Additive Modelling using the
mgcv
package. This package provides a comprehensive and powerful set of methods for modelling data.
In this case, however, the model is a relationship between time and pollutant concentration i.e. a trend.
One of the principal advantages of this approach is that the amount of smoothness in the trend is optimised in the sense that it is neither too smooth (therefore missing important features) nor too variable (perhaps fitting ‘noise’ rather than real effects).
Some background information on the use of this approach in an air quality setting can be found in Carslaw et al. (2007).
The user can select to deseasonalise the data first to provide a clearer indication of the overall trend on a monthly basis. The data are deseasonalised using the The
stl
function.
The user may also select to use bootstrap simulations to provide an alternative method of estimating the uncertainties in the trend.
In addition, the simulated estimates of uncertainty can account for autocorrelation in the residuals using a block bootstrap approach.
NO
NO
2
O
3
(Mean)
(Max)
(983435 x 39)
(23602440 x 34)
cof@qualityexcellence.es
Data allow your political judgments to be
based on fact, to the extent that numbers describe realities.
Hans Rosling
NO
NO
2
O
3
Conclusiones - I
Las concentración de NO2:
especialmente problemática en los meses de invierno.
ha crecido durante 2015, pero hubon mínimo en la serie histórica de los años 2012-2014
La concentración de O3:
está creciendo con el tiempo
y ya es especialmente preocupante en julio.
Conclusiones - II
La distribución de estaciones está permitiendo obtener resultados "optimistas" de NO2.
Faltan estaciones en zonas especialmente contaminadas (arterias principales).
Para el caso de O3:
La estación de la Casa de Campo tiene un comportamiento diferente al resto.
Sería necesario extender su medida a más estaciones dentro de la ciudad.
Se miden pocos contaminantes y no se mide el viento, empobreciendo el análisis
Conclusiones - III
Para conjuntos de este tamaño...
data.table
.
Especialmente para cálculos por grupos y transformaciones
long/wide
reshape).
Pero cuidado que para ciertos cálculos la sintaxis se complica especialmente (.
SD, .SDcols
).
Y en otros casos, más ventajoso seguir operando como "data.frame" puro.

Léete el manual de "
openair
" y reproduce estos resultados...
(Mean)
(Max)
Full transcript