The Histogram service represents parameter data values in an output visualization (a Lat-lon map or time-series) by depicting the distribution of the selected parameter data — the number of times (frequency) that values of the parameter occur — in the output visualization. The histogram will appear as a set of vertical bars where the height of each bar (and the Y-axis) indicates the number of ocurrences of parameter values in a certain range. The range values will be shown on the X-axis.
The Histogram service will operate on all parameter values in a generated Lat-lon map or time-series. The actual data used for the visualization output is contained in an HDF4 data array output file.
GIF image of the histogram bar plot
Corresponding HDF data file array with the summed occurrences of the data parameter values in each data range.
The Histogram service calculates the density distribution of a set of elements in an array. The array elements are divided such that they fall into a bin of some width between a minimum and maximum value. The histogram can be represented by the function:
where n is the total number of bins in the sample, and where xi is the ith bin whose elements have a value equal to i. In Giovanni the default number of bins is arbitrarily set to n = 50, with the minimum and maximum values set to the dynamic range of the data (unless optimal minimum and maximum values are pre-defined).
Histogram Service Types
Histogram: No averaging (calculation of mean values) is involved. The histogram is produced using all of the grid points in the selected region and time interval. This histogram type is for a single discrete time interval, which could be a 3-hour, one day, or one month period.
Histogram, Area-averaged: If a histogram of a time-series is desired, the spatial region is first averaged for each time interval, and the histogram for each time interval is produced.
Histogram, Time-averaged: If the data span multiple time intervals (i.e., hourly, 3-hour, day, or month), the values at each grid point are first averaged over time, and then the resulting histogram is derived for the selected area. This is similar to the way that a time-averaged lat-lon map is calculated, except in this case the output values will be depicted as the histogram rather than being mapped as averaged values.