predictPrice: Predict a price for an Airbnb listing using the...
In NoahZinsmeister/Rbnb: Experimental Front End for the Airbnb API

Description Usage Arguments Examples

View source: R/predictPrice.R

Given a listing ID, predictPrice uses the xgboost package to predict a price for that listing based on its characteristics and data from nearby listings. A listing ID can be found on the end of the URL for the listing on Airbnb's site. For example, a listing URL for Airbnb looks like https://www.airbnb.com/rooms/17634206?s=QAXAT6DD. In this example, the listing ID is "17634206." Passing this value as a character string will yield a price prediction based on its characteristics, implying over- or undervaluation.

1
2
3

predictPrice(listingID, maxSample = 500, nfold = 10, nrounds = 1000,
  early_stopping_rounds = 50, max_depth = 4, eta = 0.05,
  listing.detail = NULL, trainData = NULL)

`listingID`	the ID that Airbnb has assigned to the listing. This can be taken from the end of a listing's URL. This should be passed as a string (a character vector of length one).
`maxSample`	The maximum number of nearby listings on which to base the price prediction. There is a tradeoff-a higher value will increase prediction precision, but it will take longer to run.
`nfold`	For advanced users. The number of cross validation folds used by `xgboost`.
`nrounds`	For advanced users. The maximum number of rounds that `xgboost` will run when finding the optimal number of rounds through cross validation.
`early_stopping_rounds`	For advanced users. This parameter will cause xgboost to stop training new models after the specified number of rounds if cross validation error does not improve.
`max_depth`	For advanced users. The maximum depth of a tree fitted by `xgboost`.
`eta`	For advanced users. The shrinkage parameter used to control the learning rate for `xgboost`.
`listing.detail`	a dataset for the particular listing if you've already pulled it using `listingDetails`. Note that this should be a dataset of only one listing (as in `listing.detail`).
`trainData`	a dataset consisting of listings with details if you've already pulled it using `searchLocation` and `addDetails`.

data("trainData", package="Rbnb")
data("listing.detail", package="Rbnb")
price.prediction <- Rbnb::predictPrice("12170773",
       listing.detail = listing.detail, trainData = trainData[1:100])