search

DS4B-ANU / help

this is where to ask for help with your coding in BIOL3207/6207
MIT License
2 stars 0 forks source link

Request to use BOM raster data in final project #8

Open nickboffa opened 4 months ago

nickboffa commented 4 months ago

  1. A link to the dataset and a publication (or website) that explains what it is: There are many datasets, but all that I've find that are useful so far are in a link like this http://www.bom.gov.au/climate/maps/averages/temperature/?maptype=mean&period=win&region=aus You can navigate to similar maps via that link. Data can be downloaded by clicking "Grid". I potentially haven't gotten all the raster data I want to use yet, and probably want to download more from links like the above. I hope this is ok? My main dataset is still from ALA.

  2. A summary of the structure of the database and what it contains It is raster data in a .txt file. The values are whatever is specified in the map (so for the above link, it's the mean winter temperature across Australia)

  3. A sentence or two on why this dataset is both biological and interesting It's interesting because who doesn't like climate data (e.g. it vindicated my belief that Alice Springs is genuinely cold even compared to more southern places in winter http://www.bom.gov.au/climate/maps/averages/frost/?maptype=lt0deg&period=an&climatology=7605). The data isn't inherently biological, but becomes really useful when used in conjunction with biological data.

  4. A sentence or two explaining why this dataset is large and complex Initially I thought it was complex, but turns out the raster package makes it easy. It is fairly large however, with data across Australia (sometimes in 0.025 degree intervals)

  5. A summary of the types of questions and/or aims that could be addressed using this dataset (perhaps in combination with other approved datasets)

  6. The license under which the dataset is released (if you can find it) It says "Unless otherwise noted, all maps, graphs and diagrams in this page are licensed under the Creative Commons Attribution 4.0 International Licence". Whether this includes the data, I'm not sure. https://creativecommons.org/licenses/by/4.0/

DS4B-ANU commented 4 months ago

@nickboffa, all good. Can you quickly summarise what you are doing with the data though? I think from a quick glance at your repo that you are doing something with cane toads, but I'd like to know the general plan so I can give feedback. Regardless, safe to say that as long as your'e combining this with some biological data (e.g. distributions) it's all good.

nickboffa commented 4 months ago

I'm using it as the predictor data for the random forest model. So it's been given climate statistics (which is the BOM data above) of a cell (not the longitude and latitude though) and tries to predict the number of toads observed in that cell.

DS4B-ANU commented 4 months ago

Got it. Do you think number of toads is what you want? Or presence/absence? If you're looking at building a distribution map, it seems to me that presence/absence might be better, but I'm sure the paper you are following has the answer (and probably that's what you're already doing).

My concern, such as it is, is that number of toads could be really heavily biased by sampling effort, and that could mislead the model.