Data output validation is vital
We will often send multiple versions of an analysis to the client, simply because we have an updated understanding of the questions we want to answer, or new data came in, or we added projections to the structural analysis.
Never, never, NEVER send data that has changed and you didn’t know about it, because you will be blindsided by the client questions about it and they will lose faith in the analysis. - If you sent a segmented cohort named “Blue widgets” with the last analysis, and it is included in this analysis as well, it better have the same numbers as last time. If it doesn’t, the client will notice.
The golden rule of sending outputs
Always confirm that either 1. The data you’re sending now is the same as what you sent before, or 2. The data is different, and you know and can explain why (and do always explain why)
Data validation tool
It was attempted, and failed at Lucky8. 108.40.20.50 Data Analysis - Data quality tool
Graph:
- 108.40.30.10 Data Analysis - Data output validation is vital to 108.40.30 Data Analysis - Tools and environment best practices
- 108.40.30.10 Data Analysis - Data output validation is vital to 108.40.20.50 Data Analysis - Data quality tool
- 108.30.10 Data Analysis - Structural analysis to 108.40.30.10 Data Analysis - Data output validation is vital
- 108.30.20.10 Data Analysis - Predictive analysis best practices to 108.40.30.10 Data Analysis - Data output validation is vital
- 108.40.30 Data Analysis - Tools and environment best practices to 108.40.30.10 Data Analysis - Data output validation is vital