 Hello everybody, my name is Stefan Moritz, I'm from the Institute of Data Science of TH Cologne and today I'll be talking about visualization of missing data and mutations in time series. As many of you probably know, missing data is nearly everywhere. And that's also true for time series, especially in sensor data. Missing data is quite common. The reasons are usually manifold, problems with data recording, problems with data transmission or problems with the data processing. The MQTS package tries to hack it there. It features three different kind of functions. First are the imputation functions. These are functions for replacing the missing values with reasonable values. Then there are specialization functions. And then there are stats and data sets. These are functions for printing missing data stats and a benchmark data sets for comparing imputation algorithms. Actually, we have new visualizations in the Street at One version of the package, which we'll be talking shortly about. The package is currently ready on GitHub and soon will be on CRAN. Once you have a new data set, what you want to do is getting an overview over your data. Maybe you have already realized, well, there are some missing values. And now you want to investigate further about just distribution of these missing values. But what you're doing is, you use a ggplotna distribution function to get a first good impression, where are my missing values, how are they distributed in the time series and what can I maybe do there? As you can see in this plot, areas with missing values are highlighted in red and you can perfectly see where they are located in the series. This gives you a nice first overview and once you're done with this, you can even dig deeper into the missing data patterns, for example, with the ggplotna gap size function, which basically gives you a ranking of the occurrence of gap sizes. With gap sizes, we mean an ace in a row. So there could be two an ace in a row, there can be five an ace in a row. And if you have certain patterns, which seem quite unlikely, for example, if your whole data set has always six an ace in a row and no other gap lengths, then there might be something strange going on and you might want to investigate further. These plots with additional information about the patterns can be quite useful and there's also other plots than this ggplotna gap size function to do this. And once you're done with doing your first analysis, you might want to explore your imputation results. Maybe like here, you use the ggplotna underscore Kalman function, which is quite an advanced imputation function, it's Kalman smoothing on state space models. And now you have your results, PNM, and now you want to have a look how good they are actually. So you use the ggplotna imputations functions, which gives you, as you can see in this plot, a nice overview of how your imputed values fit into the time series. And as you can see here, this one fits quite good. Could already be a good solution, actually. This was a short overview over three important plots, and there's more plots to the package, which we can't show today, but what I'll show you is, the package itself is easy to use. The output is basically a ggplot object, which can be of course customized in the ggplot two-way. As you can see, you have the gap size plot, and then you make plus, see in light position equals none, and the default legend would disappear. And you can do all kinds of ggplot two magic and adapt the plot to your needs there. But actually, as we know, not everybody is a ggplot two professional, and we have also made the most important parameters available as quick adjustments. So you can also say legend equals false, that would also give you the same result as a ggplot. And you can do this for the maybe top five most important arguments, which you need for plot adjustments. So it should be also quite easy to use for not ggplot two pros. And well, that's it, feel free to contact me if you have questions or suggestions. And also, of course, take a look at the GitHub repo. And I guess there's a lot to explore and maybe a lot of useful you could use. And thank you a lot for your attention.