Threshold Data Sciences
Analysis you can defend.
I help organizations answer three questions with evidence: Did it work? Can you trust the number? What happens next?
Threshold Data Sciences is the independent practice of Timothy D. Spivey, Pensacola, Florida.
Three questions
-
Did it work?
Impact & Evaluation Analysis
I measure whether a change made a difference, how much, and how sure you can be.
Event studies · Difference-in-differences · Survival models · Equivalence testing
-
Can you trust the number?
Data & Measurement Audit
Before a number drives a decision, find out what it is actually counting.
Data profiling · Entity resolution · Metric definitions · Python and SQL
-
What happens next?
Predictive & Retention Models
Who is likely to leave, what it is worth to act, and where effort pays back.
Calibrated classification · Lifetime value · Uplift · Forecasting
A worked example
-
Demonstration · Public data · Not client work
What happens next?
Who Leaves, Who to Save, and What It Costs
A retention playbook from churn analysis to the targeting decision: calibrated models, profit-based contact rules, uplift on a randomized experiment and a fairness audit.
Calibrated classification · Profit-based targeting · Uplift modeling · Fairness audit
Public and fictional datasets; findings checked against operators' public disclosures.
Who does the work
I've spent nearly twenty years building reporting pipelines, executive scorecards and workforce forecasts in telecom operations. I have doctoral training in business analytics, an MBA in business analytics, and two articles in the ISACA Journal.
How an engagement runs
Standards
-
Reproducible code
You get the code that produced every number.
-
Documented assumptions
Every assumption and limit is stated in writing.
-
Your data stays yours
Used only for your engagement, then returned or deleted.
Latest writing
From Threshold Effects, a numbered series with sources linked in the text.
-
No. 011
Five Years to Fiber
Gulf Coast ·
BEAD is finally putting fiber in the ground around northwest Florida and south Alabama. Five years of process decided who gets to build — and it wasn't construction capability that decided.
-
No. 010
A Fair Fight with SOX 404
Regulation ·
Arguing against myself: SOX Section 404 is the strongest counterexample to this series' thesis. The argument survives — but comes out more precise than it went in.
Two ways to start
A Second-Opinion Method Review — a written memo on an existing analysis, model or vendor claim — or Paid Scoping, a one-week written analysis plan for a project that isn't clear yet. Either starts with a free 20-minute fit call.
Book a 20-minute fit call See both offers
Or email timothy@thresholddatasciences.com.