This forum is shutting down! Please post new discussions at community.naturalcapitalproject.org

Recreation model, Can we select photos?

Hi all,

Actually Im running the recreation model (3.3.1), I would like to know if I can select photos from flickr tagged like "nature", "water" "geology"...
I dont like the results that I have because there are a lot of photos from monuments, streets, and a lot of things that we dont consider "recreation service from ecosystems".

Can you help me?

My second question is about the regression model. To have an output whit regression coefficients i need a csv file, but i dont understand whit which layer it have to be related, have I to make a csv file whit LULC information about each pixel? If that dont work i can do it by mi own but i prefer understand how it works.

Thanks in advanced

Alberto

Comments

  • DaveDave Member, Administrator, NatCap Staff
    Hi @Alberto,

    Good questions. Right now we don't have the capacity to filter the photos based on tags. The only options are to filter based on location and year. So you could consider re-drawing your AOI polygon so that it excludes areas you are not interested in.

    But a better approach might be to leave your AOI the same and rely on the regression model results to indicate which attributes of the landscape are important for recreation. Most likely, things like streets and monuments are important drivers of visitation patterns, but a regression might tell you that natural landscapes are also important. Here's an example of a regression approach that might be similar to yours: https://figshare.com/articles/Geotagged_Social_Media_Reveals_Environmental_Drivers_of_Tourism_on_Jeju_Island_South_Korea/2614642

    To set up the regression:

    The csv table will include the filename (and path) to other GIS layers that the model will "count" and relate to the number of photo-user-days.

    http://data.naturalcapitalproject.org/nightly-build/invest-users-guide/html/recreation.html#data-needs
    At this link, see #6 "Predictor Table" for instructions and an example.

    To use a categorical raster like LULC as a predictor, you will need to convert it to a polygon shapefile, and create a separate shapefile for each landuse type that you wish to include as a predictor variable.

    Thanks,
    Dave
Sign In or Register to comment.