Theil-Sen Trend Analysis

[morshedahmed17]

Question

I removed some months from the dataset for the Theil-Sen trend analysis, which I do not want to show in the plot. However, those points still appear in the plots, not exactly where they were before. Is it possible to remove them completely? Also, is it possible to remove the line joining the points?

[jack-davison]

Hi,

We'd be able to best help you if you can provide a reproducible example.

TheilSen() should only plot the data you use as an input, e.g.,:

library(openair)

alldata <- mydata

TheilSen(alldata, "nox")
#> Taking bootstrap samples. Please wait.

lessdata <- selectByDate(mydata, year = 2000:2002)

TheilSen(lessdata, "nox")
#> Taking bootstrap samples. Please wait.

^{Created on 2024-02-21 with reprex v2.0.2}

More than happy to help if you can provide more info!

Jack

[morshedahmed17]

Hello Jack,

Here is an example. I removed the entire 2012 ethane data. However, the monthly values are still showing in the plot.

[mooibroekd]

When using the "deseason" option missing data is actually imputed using a Kalman filter and a Kalman smooth prior to doing the loess.

In other words, that is the probably the reason why the missing data is there. Quickest way to test is by setting "deseason" to FALSE.

[morshedahmed17]

You are right. Just one quick question.

"If I set 'deaseason = FALSE', will the resulting trendline simply be a regression line?"

[mooibroekd]

No, the output will still be a Theil-Sen analysis, but on data that has not been deseasonalized (meaning actual measurements).