Description of the Scatter Plot tool

A scatter plot is a common and effective technique used to analysis the relationship between one variable and another. The Openair Scatter Plot tool makes it possible to produce scatter plots consistent with the other Openair tools, and incorporates useful functionality such as fitting smooth or linear trends between variables, or binning or smoothing the data which can be useful where a dataset is large and there is a considerable overlap between data points.

The pollutants measured at each site are shown in the variables box once a site has been selected from the drop-down list of site names at the top of the page. The number of pollutants measured at each site varies across the network. To perform a scatter plot 2 variables need to be selected for the analysis and can be selected/de-selected by pressing and holding the CTRL key whilst highlighting the desired pollutants using the left mouse button. Modelled wind speed and wind direction from the UK air quality forecast are available in the variables list along with the pollutants and can be selected in the same way. The x and y axis variable menus become available once 2 have been selected from the Variables list allowing the choice of which variable is associated with which axis.

Data can be plotted over user defined date ranges by clicking on the drop-down lists containing the day, month and year and using these to change the start and end dates accordingly. Data can be chosen on an annual basis by highlighting the 'Specific Years' option and using the year drop-down menu to alter the start and end year.

The Type menu allows the user to split the data in several ways. The default option will produce a single plot using the entire data over the selected time range. Other types include 'Weekday' (splitting the output into plots by day of the week), 'Month' (for plots by month of the year), 'Season' (for plots split into normal seasons), and 'Wind Direction' or 'Wind Speed' (for plots split by modelled wind speed or wind direction from the UK air quality forecast respectively).

The Method menu allows the choice of producing a standard scatter plot with the 'Scatter' option where each corresponding x and y combination is presented using a point on the plot. The 'Hexbin' option is useful when the dataset is large; producing a plot where data points have been binned into hexagonal cells and the colour scale represents the count of data points within the bin. The 'Density' option will produce a smooth 2 dimensional representation of the relationship between the x and y variables, calculated using a kernel density estimate approach, again indicative of the frequency of data points at a given point.

The 'Fit Linear Trend' and 'Fit Smooth Line' tick boxes are available on this website when the 'Scatter' method is chosen. The 'Fit Linear Trend' option adds a linear model fit to the scatter plot with the equation of the line and the R2 value shown. The 'Fit Smooth Line' option will add a smooth fit to the underlying data with estimated 95% confidence intervals (grey shading) which can be extremely useful when evaluating if the relationship between variables is non-linear.

The Colours drop down menu allows the user more control over the colour scheme used in the plot when either the 'Hexbin' or 'Density' methods are selected. The clustering of data within a scatter plot can be particularly evident with these methods for example providing an indication of different air masses relating to discrete pollution episodes. A variety of colour palettes are available.