TIPS FOR USE                          METHODS                          MAKE YOUR OWN CHART


What is CTBirdTrends?
We've created CTBirdTrends to make information on the population trends of Connecticut's birds easy to find with just a few clicks from this homepage. Over time, we will add trend graphics for more species by incorporating additional data from a range of existing sources. By applying the latest statistical methods to these complex datasets and exploring the best ways to visualize them, we hope to make up-to-date information on Connecticut's bird population trends understandable and engaging. Currently, we have posted a collection of charts that allows you to view the statewide population trends for 108 of the bird species that breed in Connecticut. You can view pre-packaged charts for small groups of species or customize the chart display by selecting any combination of the 108 available species and labeling them by habitat or the statistical significance of the trend. We've also recently added an animated chart option that allows you to view the spatial variability of the raw data that goes into estimating a trend. The all-species info-graphic shows the population trends for all 108 species on a single graphic. For for information on how to interpret the charts, click "tips for use" above.

Click on a link on the left to open a pre-packaged chart or make your own.

What am I looking at?
The pre-packaged charts all show change over time in the average number of birds counted on a BBS route in Connecticut. This is labeled on the charts as "Average number of birds counted on BBS routes". To hide a species, click on the X next to its name or its trend-line. Hit refresh to revert to the full chart.

Who created this site?
This website is being developed by the UConn Ornithology Research Group.

For more information, contact Chris Field.

Recommended citation: Field, C.R., Elphick, C.S. 2012. CTBirdTrends.


Using the customizable chart
The Motion Chart has three chart types - bubble, bar, and line - which can be selected by clicking the tabs on the top right of the chart. The color scheme of the bubbles, bars, or lines can be selected using the drop-down menu below the tabs. Just refresh the page to return to its default state.

The line chart is the clearest type and the best way to view trends for larger groups of species simultaneously. The most informative view is with "Average number of birds counted on BBS routes" selected for the y axis. Select which species you would like to display from the list on the right panel and the chart will automatically rescale the y axis appropriately.

The bubble chart view allows you the animate trends over time. You can pick any variable for the x and y axes, but it is most informative when "Average number of birds counted on BBS routes" is selected for the y, and "Time" on the x (this is the default view when the page is refreshed). Once the species you are interested in viewing on the chart are selected, check "Trails" and press play to watch the animated trends. In the bubble chart view, labels for your selected species appear automatically, which can obscure parts of the trend line. You can fix this by individually clicking and dragging the labels out of the way.

The bar chart is good for viewing a summary metric of population change for a large group of (or for all) species. One interesting view is to select "Annual population size" for both the x and y axes. Choose "Habitat" for color, and click "Deselect all" to display all species. Roll your mouse over individual bars to see which species they represent.

For more detailed information, check out: http://code.google.com/apis/chart/interactive/docs/gallery/motionchart.html


The data
We used the data for Connecticut from the Breeding Bird Survey (BBS). The BBS is a continental-scale, long-term survey that is conducted by volunteer birdwatchers and coordinated by the U.S. Geological Survey. Compared with other large-scale North American monitoring programs, the BBS offers the greatest degree of spatial and temporal standardization, and therefore is most appropriate for estimating an index of population change across statewide, regional, or continental scales.

The chart
The Motion Chart being used to display the resulting trendlines was originally developed by Gapminder. The rights to the software were purchased in 2007 by Google, which recently released a free version that can be used with GoogleDocs or R (1) and embedded in any web page. The flexibility of the Motion Chart allows us to convey considerable information about the nature of changes in Connecticut's bird populations with simple, attractive graphics that are easy to use and interpret.

The model - in a nutshell
We analyzed the data using Bayesian hierarchical methods that model yearly variation around a linear trend line (Link and Sauer 2002 Ecology 83:2832-2840). The models were species-specific and based on zero-inflated Poisson regression with route and year effects to estimate for each species the average number of birds per BBS route ("Average number of birds counted on BBS routes") for each year of the survey. The habitat-wide trends on the bar chart were calculated using the methods in Sauer and Link (2002 Ecology 83:1743-1751). We conducted the analysis in WinBUGS (2) and R. For more information on the model or analysis, contact Chris Field.

Explanation of the variables
"Annual population change (%)" - the average percent increase or decrease of the population over the time period of 1966 to 2009.

"Average number of birds counted on BBS routes " - the average number of individuals seen per BBS route.

"Alpha=0.05" - S if the value for annual population change is statistically greater or less than zero (with 95% certainty); NS if it is not.

1) R Development Core Team (2008). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org.

2) Lunn, D.J., Thomas, A., Best, N., and Spiegelhalter, D. (2000) WinBUGS -- a Bayesian modeling framework: concepts, structure, and extensibility. Statistics and Computing, 10:325--337.