Time Series Scalar Plot
Time-series single-sensor plots for sensors collecting one-dimensional data are described here (i.e., every timestamped data record refers to a value at a single point in space). This definition may include instruments on mobile platforms such as Wally (a crawler in Barkley Canyon) or the Vertical Profiling System (VPS). However, gridded data from instruments like multibeam sonar and echosounders are excluded. For data searches defined by location and data source, (e.g. CTD at Folger Deep), new plots will not be made if the device contributing the data is modified, (i.e. if the CTD is swapped for another). In this case, the extent of each deployment will be shown on the plot with triangular markers and a legend. For data searches defined by instrument type, (e.g. CTD 10600), if the device is moved (e.g. from MTC to Folger Deep), new plots will be made for each location.
Revision History
- 20091208: Data product initially released
- 20100513: Subsampling option introduced
- 20101130: 1 minute interval added to subsampling options
- 20120627: station search fully supported
Parameters
Plot Types
For No Resampling and Resampling by Averaging
It is not possible to use a one plot fits all approach with such an expansive, diverse data set, therefore several plot types are used, determined by the number of data points to plot.
Line Plot with Markers:
This type is known colloquially as 'knots on a rope'. It is used for plots with greater than three data points and less than 1800 data points (30 minutes of 1 Hz data). The markers used are dots that size with the number of data points in the plot, indicating the exact time and data value of each data point. Line plots are used to illustrate the trends in the data and guide viewer's eyes through the data points sequentially. However, line plots assume a linear progression between data points, so users are cautioned not to read too much between the markers. Lines are broken to avoid interpolating through data gaps. Data gaps are defined as 1.9 times the sample period. This value was chosen for some allowance for jitter in the time stamps of the data. This is also why markers are used, to show the exact timing of the data points.
Scatter Plot:
For less than 3 data points or 1801 to 14399 data points (30 minutes to 4 hours of data at 1 Hz sampling), a scatter plot is used. The markers used in the scatter plot are the same as the line plot above. In a line plot, as the number of data points increases, it becomes increasingly difficult and potentially confusing to see the line. In a scatter plot, as the number of data points increases, the markers start overlapping to great degree, potentially masking trends and distributions within the data.
Heat Map:
In this example of Folger Pinnacle CTD pressure data, one can see the tidal cycle, plus the variation of the pressure distribution due to the sea state (there was a moderate-size west swell running between the 8th and 10th of June, 2013).
When the number of data points to be plotted is equal to or greater than 14400 (4 hours of data for 1 Hz sampling), it becomes more difficult to resolve the data points in scatter and line plots, so plot generation defaults to a heat map. (To avoid heat maps, for 1 Hz data, select 1 minute resampling for a search duration of up to 10 days, select 1 hour resampling from 10 to 600 days, 1 day resample thereafter. Use averaged or min/max resampling.) A heat map is like a scatter plot where the user can discern the relative number of markers at a location where they are overlapping. Therefore, it resolves the overlap issue in scatter plots of large numbers of data. Users are able to see the distribution and trends within very large time series data sets. For advanced users, the matlab code that makes a heat map from ONC mat files is: z = hist3(data.time, data.dat); imagesc(z);. A heat map is a 3-D histogram, represented as an image, coloured by the number of data points that fall in each time by value bin. Colours range from white (no data), to light blue (least amount of data), to dark blue (greatest amount of data). Bins are sized to have at least 400 along the x axis, 300 along the y axis, so as to offer as high a resolution as a scatter plot. The actual number of data points in each bin is not shown as it is not comparable from plot to plot, due to the variable sizing of the bins. The relative number in each bin is shown, this is called the occurrence frequency.
For Min/Max Resampling
Currently, the min/max plots consist of the two lines, one for each of the maximum and minimum data points. The next version will fill in the space between the lines. The lines are staircase or step-function plots, so when filling in the space, the plot will effectively become a bar plot, similar to plotting utility, but with the added benefits of higher resolution, having the QAQC flags plotted on the raw data, markers for the device deployments and the ability to control the resample period. Here is the pre-release version:
Formats
Plots are available in PNG and PDF format. Basic metadata information is included in the titles. Note that the logo in the PDF product may appear 'fuzzy' in some viewers, see the PDF page for more details.
PDF example:
FolgerPassage-FolgerPinnacle_CTD_Pressure_20130601T000028Z_20130601T000458Z.pdf
Discussion
To comment on this product, click Add Comment below.