Tag: data-research
Concepts
- Aggregating Raw Records Into a Company Panel
- Building Your Own Vintage Archive
- Coverage Analysis Across Your Universe
- Designing a Vendor Data Trial
- Entity Resolution When Identifiers Do Not Match
- Extracting a Signal From Unstructured Records
- Getting a Useful Trial Sample
- How Many People Already Have This Dataset?
- How Much History Do You Actually Need?
- Evaluating a New Dataset: The Full Checklist
- Keeping a Data Inventory Researchers Will Use
- Mapping a Vendor Dataset to Your Universe
- Marginal Value Over Data You Already Own
- Matching Data Frequency to Signal Frequency
- Matching Filters, Screens and Exclusions
- Normalising Counts Across Entities of Different Size
- Onboarding a Bought Dataset Into Production
- Panel Churn and Composition Drift
- Panel, Cross-Section or Event Stream?
- PII and Anonymisation Checks Before Onboarding
- Questions to Ask a Data Vendor
- Reading a Vendor's Own Backtest
- Reconstructing a Paper's Data and Universe
- Renewing or Cancelling a Data Subscription
- Revision Analysis: How Much Does History Change?
- Sample Size Per Name and Signal Reliability
- Seasonality and Day-of-Week Effects in Raw Feeds
- Signals From Sparse, Intermittent Data
- Building the Cost-Benefit Case for a Data Purchase
- Timeliness: When Does the Data Actually Arrive?
- Why Replications Fail
- Your First Hour With a New Dataset