Graphed up results from both velocity data and area data, ready for analysis. Some are looking pretty crazy … and might need replotting with new axis. (Go look here: http://www.harding.edu/fmccown/R/)
Worked out a plan to automate APNG creation, which I will try to pull together when I have a spare 2 hours.
Bunch of stuff today:
Today was furthering graphing work with R, where I developed a script that will pump out boxplots like mad for the different experiments. All quite cool, but now need to start analysing them — e.g. print them off and have a look at them all. (Measure gradients, etc). Also need to do the same for the velocity measures.
More links of use:
See here: http://www.ats.ucla.edu/stat/splus/faq/subset_R.htm
Note that you have to use == for conditional equals, e.g y== 2
Random link that could prove of use: http://www.fort.usgs.gov/BRDScience/learnR9.htm — Regression diagnostics
This will probably prove to be useful: http://cran.r-project.org/doc/contrib/Lemon-kickstart/kr_scrpt.html — how to write scripts in R.
Worked out why I had 10 pieces of info in the velocity output: it’s a velocity for each bird. Ergo, experiments with 50 birds x 10 results = 500 columns of data. => not sure how to graph and visualise.
Started looking at generating base data for area today. Option a/ do it by hand in Excel, or b/ re-jig code. Going with option b/ but there is a gremlin that needs extracting from the works.
NB 2 way lookups in Excel? See here: http://skp.mvps.org/xl00002.htm. V. good, and means can get results into timestep, measure, grouping quite quickly.
Changing the code to not calculate stdev of velocities. In theory the change made should have worked … but why when I have 5 birds I get 10 results??
This book may be of use:
By Elizabeth Spencer Allman, John Anthony Rhodes
Some links that may be of use: lectures on model fitting http://www-users.york.ac.uk/~pjh503/mathematical_model/ and some userful terms for additional reading around on here: http://isolatium.uhh.hawaii.edu/m205L/lab6/fit/fit.htm
Plotted graphs of area and velocity against the individual parameters tonight. In short, some of the velocity results are linear. It’s fair to say the area results are not, and don’t look to be easily transformed …
Completed weekly planning session. Had a bit of a tidy up re: data so have a ~80MB Zip file with mostly all of my summarised data in it — where has some of the random data for velocities gone?
Then started on charting the data, independent variables vs. area to begin with.
