The goal of farrell is to provide an interactive interface to Data Envelopment Analysis modeling in R. The farrell package is built upon Benchmarking.
You can install the development version of farrell with:
You can run:
or if you’re working on RStudio, just click on Addins then farrell.
Hit Browse… to upload your data frame in a csv format. All the inputs and outputs must be contained within the data frame (each column for each input/output). Further, the data frame needs to contain an identification column in order to identify Decision Making Units. It can be a numeric or a character column.
In the following examples, we use the mtcars data frame which has been exported in a csv format with an additional column: cars name.
Within the Model Tuning tab, you will select the input and output variables, then you determine your identification column. Then you choose the Returns to Scale assumption between: crs, vrs, irs, drs, add and fdh. After that, you determine the orientation of the model, whether input or output. Finally, hit Calculate Efficiency to get the results.
Let’s for example consider mpg and disp as the output variables and wt as input. We choose cars name as the identification column and model an input-oriented model with crs assumption.
The Efficiency Results tab displays the efficiency scores along with the peers for each unit in a descending order. You have the ability to download the result in a csv format. The tab also provides a summary of the distribution of the efficiency scores.
In the Lambdas tab, you get the contribution of the peers to the inefficient units’ score.
The SE tab provides the Scale Efficiency score of each DMU under consideration.
The Slacks tab displays a data frame containing the sum of the slacks and the slacks for each input/output variables.