What you will learn in SGDS2:

Stats, Graphs, and Data Science², continues on the foundations of SGDS¹ by building on the learning objectives in greater depth, breadth, and detail.  We continue examination of the theory, concepts, and practice of data science as applied to asset analytics.  The learning follows the development sequence of valuation and risk analytics.  Starting from the problem identification (scope of work) in measurable terms, we proceed to:

• Identify the overall project data frame;
• Define the CMS^© (Competitive Market Segment), the ideal data set;
• Resolve the bias-variance tradeoff to optimize accuracy and precision of results;
• Perform necessary data wrangling (analogical to traditional confirmation/verification)
• Portray data visually
• Reduce to the CMS^©
• Detect quantify missing/needed predictor variables
• Create new calculated fields
• Transform functional relationships (e.g., linear to logarithmic)
• Create Transformed data levels
• Identify and handle outliers
• Recognize needed predictive confirmation/verification

This class is substantially experienced in R, RStudio, and R packages, including ggplot2 and tidyverse.  A comparison and integration of Excel® is provided for workflow efficiency.  Methods of transition from canned appraisal software and the spreadsheet— to a real analytics package is reviewed.

Specific analytics include non-linear price indexing, seam regression, introduction to GIS in R, Data selection algorithms (particularly hierarchical cluster analysis), sources of help for R.  Also resolution of conflicts in USPAP, AI-SVP, IVS, and EBV^© (Evidence Based Valuation) practice, and reporting in a data-stream delivery through dashboard views.

In addition to George Dell, SRA, MAI, ASA, CRE, as your instructor –

Bruce Hahn, MAI, SRA, CRE, will be a second instructor in the class.

Bruce is an independent fee appraiser, applying R analytics to residential and other assignments.