Please download the Spatial Interpolation Data Function for Spotfire® from the exchange.
Release 2.0.0 onwards
Overview
This Data Function interpolates a continuous variable across a 2D region by performing either LOESS (LOcal regrESSion) or IDW (Inverse Distance Weighted) on the latitude and longitude coordinates, to best predict a target. The download contains a data function you can import into your own .dxp and a template you can bring your own data into.
The two interpolation methods capture local structures in different ways.

Inverse Distance Weighted (IDW) is a spatial interpolation method used to estimate values at new locations, based on known values observed at a set of locations.
 The power parameter (inverse distance weighting power or idp) in the IDW formula lets users control the rate at which the influence of points diminishes with distance. The higher the value of this parameter, the more influence is assigned to nearby points.

LOESS fits a polynomial surface to the data. Fitting is done locally using points in a neighbourhood of the target point, weighted by their distance from the point.
 The degree of the polynomial is controlled by the degree parameter, and size of the neighbourhood is controlled by the smoothing parameter. Larger values of the smoothing parameter produce results that are less sensitive to local noise in the data.
Details about the data function
Techniques used: Inverse Distance Weighted, LOESS
Required Packages: data.table, sf, stars, gstat, stats
Description: This data function interpolates a variable across the twodimensional latitude and longitude coordinate space. See Readme.pdf in this distribution for details on inputs and outputs. In order to run the interpolation, you will need to provide a dataset with longitude/latitude coordinates and one or more columns representing the values to interpolate. The included sample dxp contains an additional utility data function (Create Grid) that can be used to generate a test grid of locations to apply the interpolation to. The main data function (Spatial Interpolation) uses the generated grid as target table, but users can easily bypass the grid creation and upload their chosen target table.
Releases 1.0.0 and 1.1.0
Overview
This Data Function interpolates a continuous variable across a 2D region by performing LOESS (LOcal regrESSion) on the interaction between latitude and longitude coordinates to best predict a target. The download contains a data function you can import into your own .dxp and a template you can bring your own data into.
Details about the data function
Technique used: Local Regression
Required Packages: stats
Description: This data function interpolates a variable across twodimensional latitude and longitude coordinate space.
Inputs:

fit_longitude: Xvalues (or longitude) used for the interpolation

fit_latitude: Yvalues (or latitude) used for the interpolation

fit_response: Zvalues (i.e. your variable of interest) used for the interpolation. Should be from the same source table as X and Y values
 predict_longitude: Xvalues (or longitude) that will be interpolated to

predict_latitude: Yvalues (or latitude) that will be interpolated to
 smoothing_parameter: The smoothing level of the loess smoother used to interpolate values. Default = 0.5
Outputs:

loess_prediction: Column of data resulting from the interpolation. To be added as a Calculated Column to the Target Table
 loess_fitted: Column of data resulting from the interpolation; represents prediction at the source data for comparison. To be added as a Calculated Column to the Source Table

loess_residuals: Column of data expressing the residuals (distance from the predicted 'Fitted' values and the actual 'Source' values). To be added as a Calculated Column to the Source Table.
Reference Info
 For an overall overview of the Geo Analytics or Location Analytics capabilities of Spotfire please also review this page.
 Useful Geospatial Data Functions for Spotfire
 Below is a TAF video about this topic.
Recommended Comments
There are no comments to display.