SOFIA: Plots

SOFIA Science Data Archive: Plots

Plots (sometimes called charts) are made from Tables, and, more commonly (or usefully) from Catalogs. Plotting is covered in this section. The Tables section discusses tables more generally, and the specific case of loading catalogs is in another section. If your table has RA and Dec in it, the Visualization section covers how the catalog can be overlaid on images. If your plot is a spectrum, the Spectra section also covers that.

Contents of page/chapter:
+Default Plot
+Plot Format: A First Look
+Plot Navigation
+Plot Linking
+Changing What is Plotted
+Plotting Manipulated Columns
+Restricting What is Plotted
+Overplotting
+Adding Plots

Default Plot

When doing a position search for a fixed target, the default plot is often somewhat less than useful -- it grabs the requested positions from the list of observations, or even just the coordinates of the reference pixel (the CRVAL 1 and 2) of the retrieved images (see first screenshot below). If you have constrained your search results to only return the most centered image, this is an even more phenomenally uninteresting plot, because it has a single point. If you do a search on a moving target, the default plot is the RA and Dec of the object that it finds with data, so moderately more interesting (see second screenshot below).

This is a plot of positions of a list of EXES observations returned by a search on M42.

Positions of a list of images returned by a moving target search (left), and (right) the slightly more exciting plot of positions calculated for said moving target.

You can change what is plotted (see below) but plots of image metadata may not be what is most desirable.

Try loading in a catalog and keep reading! The default plot for a catalog will also be RA and Dec.

Note that all of these RA/Dec plots follow astronomical convention -- RA increases to the left.

To obtain a full-screen view of your plot, click on the expand icon in the upper right of the window pane: . To return to the prior view, click the "Close" arrow in the upper left.

Plot format: A First Look

If you have loaded a catalog with many (> 5,000) points, you may have an RA/Dec plot that looks something like the one on the left here. If you have loaded a catalog with few (< 5,000) points, you will have an RA/Dec plot that looks more like the one on the right here.

The difference between them is that, for larger catalogs (left), the plot is binned -- more points are encompassed in a black tile and fewer points are encompassed in a white tile. In the context of this tool, this is called a heatmap. The shades of grey correspond to how many points are encompassed in each 'cell', with the density scale given on the right hand side of the plot. For smaller catalogs (right), each individual point is shown as a blue dot. In the context of this tool, this is called a scatter plot. Note that even when individual points are shown, where the points overlap, the color is darker.

In either case, letting your mouse hover over a point tells you the values of the point under your cursor, and (if binned) how many points are represented:
for binned plots, and
for just one point.
Clicking (in an unbinned plot) highlights that point, and it stays highlighted, though you must keep your mouse on the point in order to see the information about it.

The reason the tool makes a heatmap for large catalogs is to more fairly represent the point density -- and to make the plotting faster. In these cases, though, it will not give you the option to overplot errors (see below). If you have a heatmap and want a scatter plot by default, you need to filter or otherwise restrict the catalog to have fewer points (see below). You can change the bin size and shading via the plot options pop-up (more on this below).

Plot Navigation

The top of the plot window has a row of icons something like this:

which we now describe.

Add new plot

You may or may not have this icon. Clicking on this icon adds a new plot. This has a separate section below.

Plot mode

This trio of icons controls the plot interaction 'mode'. By default, you are in 'selection' mode, as seen here -- the last icon is darker, like a pushed-in button. To activate the other modes, click on the other icons, and they become darker or "pushed in."

Zoom mode: When this mode is active, when you click and drag in the plot, the plot is zoomed to the region you have selected. Even when this mode isn't active, you can also zoom using your scroll feature on your mouse. To return to the original view, click on .
Pan mode: When this mode is active, when you click and drag in the plot, it moves around in response to where you drag. To return to the original view, click on .
Select mode: When this mode is active, when you click and drag in the plot, you are given additional options at the top of the plot : The checkmark means "select" and the funnel means "filter." The difference is that filtering (temporarily) limits what is shown in the plot, catalog, and image (see general information on filters), and selecting just highlights the points enclosed within your selection. To cancel either one, click on cancel filters or cancel selection .

Re-scale plot

Return to the view that optimizes the range of x and y to show the currently displayed points.

⚠ Tips and Troubleshooting: Did you accidently zoom in the plot with your magic mouse or touchpad? Click on this icon to reset the plot.

Save plot

Save the plot. It will save as a png file, wherever your browser is configured to save files. The saved png is the same size as it is on your screen. If you want a big version, make the desired plot big on your screen (expand the view to take up as much space as possible) before saving the png.

Undo

Restore everything to the defaults. If you've played a lot with the plot, you may want to undo everything you've done. Click this icon to restore everything back to the defaults.

Filter from plot

Pull up interactive filters. This button brings up filters for the displayed catalog in an interface like all the other tables here, except you don't see the values in the catalog themselves; you can enter filters here in the same way you can everywhere else in this tool (see general information on filters).

Configure plot

Click on this icon to change what is plotted (much more on this below).

Expand plot

Click on this icon to make the plot take up the whole browser window. To return to the prior view, click the "Close" arrow in the upper left.

Help

This icon may not appear, but if it does, it is a context-sensitive help marker, which should bring you to this online help.

Plot Linking: Plots are linked to catalog and image(s)

If you move your mouse over any of the points in the plot, you will get a pop-up telling you the values corresponding to the point under your cursor. For scatter plots, if you click on any of the points, the object(s) corresponding to that point will be highlighted in the overlays in the images shown, and highlighted in the catalog table. This works the other way too -- click on a row in the catalog, or an object in the images, and the object will be highlighted in the plot or the catalog or the image.

Changing What is Plotted

To change what is plotted, click on the gear icon in the upper right of the plot window pane:

. Configuration options then appear; the options are a little different depending on whether the points are binned or not. This section describes how to change what is plotted, i.e., the "Modify Trace" option at the top of both of these pop-ups. The overplotting option (and, for that matter, adding plots) are covered in more detail below.

This is the configuration window for a binned (a.k.a. heatmap and/or greyscale) plot. By default, the "chart options" may be hidden; to reveal them, click on the name "Chart Options" or the disclosure arrow on the right. To hide them again, click on the disclosure arrow on the right.

The configuration window for a plot that shows individual points, once fully extended, is much longer (and scrollable), and so is shown here in two parts. Both the "Trace Options" and "Chart Options" may be hidden by default; to reveal them, click on the name or the disclosure arrow on the right. To hide them again, click on the disclosure arrow on the right.

Options found in both kinds of plots

In either case, you can specify what should be plotted on each axis. The magnifying glass is a link that brings up a table that lists all of the available columns in the catalog. Alternatively, you can just start typing, and viable options appear below the box. Whatever you put in the box must match the column name as shown in the catalog exactly.

Click on the black triangle to reveal additional options.

In both of the examples above, RA is plotted on the x-axis. It has pulled the column name for the label; in this table, the column is "ra" rather than "RA", and it is case-sensitive. It has copied over the units ("deg") from the catalog, and plotted the x-axis increasing to the left as per astronomical convention. You can change what column is plotted, and whether or not errors are shown. Under "Chart Options", you can specify:

title of the plot;
labels on the x and y-axis;
whether or not there is a grid shown;
whether or not the axis is reversed (as for ra in the examples above);
whether the x-axis is on the top or bottom and the y-axis is on the left or right;
whether or not the axis is logarithmic;
the maximum and minimum values of the plot range;
the aspect ratio of the plot (e.g., square or rectangular).

By default, the boundaries of the plot are set to encompass the full data range. Here you can change the boundaries to specific numbers. (This can also be set via filtering from the plot; see below.)

You can enter simple mathematical relations in these boxes too, such as (for a WISE catalog) "w1mpro-w4mpro" to put [W1]-[W4] on one axis. Supported operators:

+,-,*,/
abs(x), acos(x), asin(x), atan(x), atan2(x), ceil(x), cos(x), exp(x), floor(x), lg(x), ln(x), log10(x), log(x), power(x,y), round(x), sin(x), sqrt(x), tan(x)
degree(x) and radians(x) are also supported -- these are the same functions as in ADQL and convert radians to degrees or degrees to radians. For small astrometric offsets, you could make a scatterplot of dec2-dec1 vs. (ra2-ra1)*cos(radians(dec1)) instead of typing cos(dec1*pi()/180). (NB: pi() is also a supported function you can use, instead of typing 3.14159.)
Non-alphanumeric column names (e.g., those with - or + or similar characters) should be quoted in expressions.

Click "Apply" to apply, and "Close" to return to the plot without making changes. (For the latter, you can also click the 'x' in the upper right.)

Options found only in binned plots

(Plots are binned by default if there are > 5,000 points in the catalog.) From the pop-up, you can control the color table that is used (greyscale is the default; there are many other choices in the drop-down menu), as well as the number of bins in the x and y directions. The default value for the number of bins is 100 in both directions.

Options found only in plots showing individual points

You can add errors. Toggle the error switch, and then additional choices appear. From there, you can select symmetric or asymmetric errors, and then you can specify an error as either an existing column in the catalog, or calculated from a column in the catalog.

Under "Trace Style," you can control whether the points are shown as individual points, connected points, or just lines connecting the points.

Under Trace Options, you have many choices.

Choose the symbol type: circle (default), open circle, square, open square, diamond, open diamond, cross, x, upward-pointing triangle, hexagon, or star.
Choose the color. By default, the point color is a mid-range blue that is darker where more points. This is specified by the rgba vector shown in the example here (31, 119, 180, 50) where the last number is in units of fraction of 1, so 0.5=50% in this example. Click on the magnifying glass to bring up a color picker window:

From here, you can click on your desired color in the top colorful box. Immediately below that box, you can change the color and saturation of the top box so that you can select from a different range of colors. Below that, you can enter numerical hex codes or RGBA values (where the value for RGB is between 0 and 255, and A is in units of percent, e.g., 50 = 50%). Finally, you can also select from a pre-defined set of 15 colors by clicking on any of the small boxes. Note that the numerical codes update as you select different colors. Click "OK" to implement your color choice, or click 'x' in the upper right to close the window without changing the color.
⚠ Tips and Troubleshooting: Don't like the transparency feature of the points that makes them darker when there are more points? Set the last value of the vector (A) to 1. Don't like the blue? Pick a diferent color entirely. Want the faintest point to be brighter than it is by default? Set the last element of the color vector ("A") to be 0.7 or 0.8.
Choose the color map. By default, all of the points are the same color, but darker where there are more points. You can change this such that the color scale of the points is tied to a column value, such as w1snr (WISE-1 signal-to-noise ratio) in a WISE catalog. If you select this option, you can also change the color scale to any of many different options (see the drop-down). Simple mathematical relations (as above) are also permitted in this box.
Choose the size map. By default, all of the points are the same size. You can change this such that the color scale of the points is tied to a column value, such as w1snr (WISE-1 signal-to-noise ratio) in a WISE catalog. Simple mathematical relations (as above) are also permitted in this box.

Example: Load a large WISE catalog. Plot w1snr (WISE-1 signal-to-noise ratio) vs. w1mpro (WISE-1 profile fitted magnitude). It defaults to a heatmap. Change the labels, making the y-axis label "WISE-1 SNR" rather than the more cryptic column header "w1snr". Change the x-axis label to "[W1]." Change the greyscale to yellow-green-blue ("YlGnBu") to make it easier to see the lowest-populated bins. Depending on your catalog, you may need to adjust the ranges. Obtain this plot:

Example: Load either a smaller WISE catalog, or the same large WISE catalog, but filter it down such that w1snr, w2snr, and w3snr are all greater than 10, which limits the number of points to be <5,000. Plot w1snr vs. w1mpro. It shows the points individually. Change the labels. Change the point color map to scale with w2mpro (WISE-2 profile fitted magnitude). Change the point size map to scale with w4snr (WISE-4 signal-to-noise). Obtain this plot:

Plotting Manipulated Columns

You can choose a single column to plot against another column, as above. However, you can also do simple mathematical manipulations.

For example, if you have loaded a WISE catalog, you can plot [W1]-[W2] vs. [W3]-[W4]. In terms of the names of the columns in the database, this is w1mpro-w2mpro vs. w3mpro-w4mpro.

If you have few enough points that the plot is not binned, you can add errors that you calculate. Here, the expression for the x-axis errors is sqrt(power(w1sigmpro,2)+power(w2sigmpro,2)) and for the y-axis errors, it is sqrt(power(w3sigmpro,2)+power(w4sigmpro,2)) -- that is, the errors for the individual photometric points added in quadrature.

Restricting What is Plotted (from the plot)

You can also restrict what data are plotted in any of several different ways.

You can filter the catalog from the table itself (discussed in another section).

You can set axis limits on the plot itself from the plot options pop-up (discussed above).

However, and perhaps more powerfully, you can set limits from the plot itself using a rubber band zoom. Click on the select icon in the plot Then, click and drag in a sub-region of the plot. New icons appear: If you click on the funnel icon, only those data points that pass the filter are shown in the plot, in the table, and/or overlaid on the image(s). (This is the behavior of 'filter', as opposed to 'select'; the former restricts what is shown, the latter just highlights the points.) For more on filters, see the filtering discussion in the tables section.

Example: Obtain a WISE catalog of a star-forming region, say IC1396. Filter down the catalog to only have detections at all four WISE bands. (Limits have undefined errors, so ask the catalog to filter down such that w1sigmpro>0, w2sigmpro>0, w3sigmpro>0, and w4sigmpro>0). Plot w1mpro-w4mpro on the x-axis, and w1mpro on the y-axis. Reverse the y-axis to put bright objects at the top. Click and drag in the plot to select the bright and red objects, and filter them down to get a subset of bright and red sources. For clarity, the screenshot here has the sources selected, not filtered.

Overplotting

At the top of the pop-up that you get when you click on the gears, you have two radio buttons:

They are "Overplot New Trace" and "Modify Trace." Modifying traces (plots) has been covered above; in this section, we will cover overplotting. This is sometimes called "multi-trace," meaning that more than one thing is plotted.

When you select "Overplot New Trace," you get a new interface that is very similar to the original interface where you selected what to plot:

As before, you need to :

select a plot type (scatter, heatmap, histogram);
tell it what column(s) (and and manipulations thereof) you want for x, y, and associated errors;
select the trace style (points, connected points, lines);
set any additional trace options;
set any additional chart options.

The best way to explain how to use this feature is probably via an example. We have a plot of [W1] vs. [W1-W4] from above. Now add on top of it a plot of [W2] vs. [W1-W4]. Click on the gears to bring up the pop-up. Select "Overplot New Trace." Enter "w1mpro-w4mpro" for x and "w2mpro" for y. Expand "Chart Options." Note that it has preserved the overall chart title from before, but has erased the X and Y labels (and lost the reversal of the y axis) because the overplot could literally be anything, and need not be the same columns or even the same units as what is already plotted. Type them in again. Here is the configuration window right before clicking "ok", and the resultant plot.

After you add the overplot, if you click on the gears again, note that the choices at the top of the window have changed. You can add another overplotted trace, modify a trace, or remove the active trace. Each trace that you add is a new 'layer' on the plot. The drop-down menu near the top of the window controls which trace is 'active' for setting the x, y, errors, trace style, name, symbol, color, etc. there is now a drop-down menu at the top of the plot: There is a legend on the plot specifying which color corresponds to which trace. In this example, the plot above has appeared using a blue and green color scheme, which may be too hard to differentiate. To change the new points' color, click on the gears, ensure "Modify Trace" is selected, select "trace 1" (as opposed to "trace 0", the first one you loaded), go down and expand the "Trace Options" and pick a different color. You can also change the legend name from "Trace 1" to, in this case, "[W2]". Click "apply" to apply the changes to the plot. Note that once you change the trace name, the relevant drop-down menus in the pop-up window and the legends on the plot update accordingly.

Note that the pop-up spawned by clicking the gears now has an additional option at the top: "Add New Chart", "Overplot New Trace", "Modify Trace", and "Remove Active Trace." From here, you can modify a trace you have already plotted (as described above), overplot another trace (also as described above), or remove the selected trace:

⚠ Tips and Troubleshooting

Right now, the overplotting only works from the same catalog -- that is, you cannot plot [W1] vs. [W1]-[W4] from one catalog and overplot [W1] vs. [W1]-[W4] from another catalog. (We enthusiastically await this capability too.)
You can easily get yourself into a physically nonsensical situation, say, by overplotting a histogram onto a scatter plot. If you find yourself in a hopeless mess, click on the "undo" icon to reset everything and try again.
When you have more than one thing (trace) plotted, double click on the legend to bring that trace to the foreground and temporarily hide the other traces.
You can overplot a scatter plot on top of a heatmap if you really want to!

Adding Plots

Clicking on this icon brings up a dialog from which you can choose to make another scatter plot (left below), a heatmap (center below), or a histogram (right below):

The options for these plots here are very similar to what is described above. You can specify which columns to plot or manipulate and plot, specify labels, etc.

Scatter plots allow you to choose points, connected points, or lines; you can add errors to each point.

Heatmap plots are binned scatter plots; you can choose what color scale and how many bins to use.

Histogram plots allow you to choose how many bins or the bin width. Note that, if you provide a minimum number, the binning starts at the minimum value you provide, and may exceed the maximum you entered in order to fit in a whole bin.

You can change what is plotted after plotting by clicking on the gears, as described above.

You can have many plots up at the same time.

You can view multiple plots all at once or one at a time by clicking on the corresponding icons above the plots (just as when you have multiple images loaded). The single box means "one at a time", the set of four boxes means "all the plots at once". If you are viewing one at a time and have more than one plot loaded, you will see the ">" and "<" signs (as in the image here), and you can scroll among the plots by clicking on these arrows (just as when you have multiple images loaded).

⚠ Tips and Troubleshooting

Note that many plots of a large catalog may make your browser run slowly.
You can force the tool to make a scatter plot even if you have a catalog of >5,000 points; it just might make your browser slow down. You can even overplot a scatter plot on top of a heatmap if you really want to!
To remove a plot, click on the 'x' in the upper right corner of the plot.