The data board has two main tasks. The first one is to offer you direct access to your experimental data. Its second task is to make your data easier to work with or even analyze it for you.

Table of Contents

  1. Getting started
  2. Overview
  3. Analysis apps
    1. Germination
    2. Growth

Getting Started

When starting the analysis board you will start in the overview section. The overview is actually used to select and pre-edit the data for all further analysis apps. This is done in two main steps which the progress bar will show when you enter the Analysis board.

The first step is to select an experiment that you would like to use for the analysis. Simply use the drop-down list to select an experiment or start typing the name of the experiment that you would like to analyze. If you made your select you will see an overview again of the experiment that will show how many blocks have been used in the experiment, how many units these blocks contain, how many sample groups (unique combinations of genotype and treatment) were created as well as the sample size (number of units within the same sample group i.e. replicates). If this is is the experiment you want to analyze simply click on step 2 in the progress bar.

After clicking on step 2 you will be presented with a plot on the main canvas that shows a specific parameter for the experiment selected over time similar to the plot below. The parameter to display can be selected in the drop-down above the plot. 

What you might realize by looking at the plot below is that there is by default only one data point per day for each sample group (unique combination of genotype and treatment). This is because there is a aggregation and grouping function included in the current view that you have selected. Before going into details about these features be aware that all filters, aggregations or groupings you apply to the data here will be used in all following analysis apps you might start. To see what options we have to play with the data open the settings bar on the right hand side.

Filtering will reduce the data set by either limiting the time for analysis or by excluding certain genotypes or treatments from the dataset. By default the time filters will be set to the start and stop date of the experiment. If you wish to further reduce your analysis (e.g. to the first days of the experiment to evaluate germination) simply select the start and end dates using the date picker. You will see that the graph will show two dashed lines that mark ANALYSIS START and ANALYSIS END (see above screenshot from May 07 till May 25). 

If you want to limit your analysis to a certain subsection of genotypes or treatments simply type their names in the two fields in the filter section of the settings menu. As soon as you start typing a suggestion of genotypes or treatments that match your input will be presented. If there is no value in the filter fields all genotypes and treatments from the experiment will be used for further analysis. 

Grouping allows combining data points within one sample group into one graph. Grouping can be used to average and hence also simplify the result of multiple replicates of the same sample group into one graph. Imagine an experiment with 5 genotypes and 2 treatments (10 sample groups) and a sample size of 10. Without grouping you would have 100 graphs to display. If you group by genotype and treatment you get only 10 graphs.

When grouping you can also decide how to group in the settings bar. By default the grouping will be defined by sample groups meaning that the group by field will show treatment and genotype. Hence, one group for each unique combination of both will be created. When only selecting genotype all treatments of the same genotype will be combined. When only selecting treatment, all genotypes of the same treatment will be combined. 

Aggregation allows combining multiple data points that are within a certain time range into one data point.  Aggregation can be used either when measurements have been taken with high frequency and should be simplified into one time point or when measurements for different samples of the same group have been taken at a different time and the time points should be normalized for all samples. Imagine a system that scans 120 plants from 12:00 - 14:00. The time point for each measurement would differ by one minute. By setting an aggregation to 13:00 you could normalize all measurements to that time for further analysis.   

When aggregating data you can decide on what time points of each day an aggregation should be made. You can define up to 24 time points (1 per full hour). One typical application would be to define one aggregation point at night (e.g. 0:00) and one at daytime (e.g. 12:00).

After selecting the time points, there are different ways how data points are assigned to the defined aggregation points. The default mode is nearest where each time point is assigned to the nearest aggregation point. as an example let's take the aggregation points 0:00 and 12:00 mentioned above. A point at 10:00 would be assigned to 12:00 because it is only 2 hours away whereas 0:00 is 10 hours away. If we switch the assignment to after only data points after the aggregation point can be assigned. In this case 10:00 would be assigned to 0:00 as it is after this point but before 12:00. 

In addition, you can select a maximum range for assigning data points to aggregation points. If we take the example above and set the range to 6h the point at 10:00 would be discarded. It is after 0:00 but more than 6h from this aggregation point. More specific the combination "after" with a range of 6h and aggregation points 0:00 and 12:00 would only assign points from 0:00 - 06:00 to aggregation point 0:00 and all points from 12:00 - 18:00 to aggregation point 12:00. All other points would be discarded. 

Last, you can select how the assigned points will be combined in the aggregated point by selecting the aggregation method. The default method is Median which will take the median value of all data points assigned to an aggregation point for this specific aggregation point. Alternative methods are the average, maximum or minimum data point.  

Quick Analysis

After making your selection on filtering, grouping and aggregation your plot will show the growth curves for your selection. There are still some plotting options to discover in the settings bar such as switching from absolute values to relative, or to limit the number of lines in your graph. Of course you can also still change your analysis settings or change the parameter you want to display. Nevertheless, if you have a good setting you should already get an impression of your data by the visualized graph. 

If there are outliers or interesting features you would like to understand you can click on a single point in the graph. The canvas will show a data table below the plot. This data table shows what samples and time points have been aggregated and grouped in this specific point that you clicked on. It is important to understand that for all further analysis the data we work on will be based on the grouping and filtering done here. 

Next steps

If you are happy with the outcome simply open any of your other analysis apps and the dataset you pre-defined here with all it's filters, groups and aggregations will be exported into the analysis app.