Phase 3 - Analyze
Objective
Perform data analysis in alignment with the defined objectives and questions.
Key Activities
- 8.0 Conduct Initial Exploratory Data Analysis (EDA)
- 8.1 Calculate descriptive statistics
- 8.2 Create initial visualizations (histograms, scatter plots, box plots)
- 8.3 Conduct correlation analysis
- 8.4 Identify potential patterns, anomalies, or areas of interest
- 8.5 Document initial observations and hypotheses generated from EDA
- Refine initial questions and hypotheses 2.1 Formulate initial questions and hypotheses
- 8.6 Address newly identified gaps or additional data cleansing activities: 6.0 Data Validation and Cleansing
- 9.0 Develop and Execute Test Scripts and Queries
- 9.1 Write scripts or queries to implement the planned analyses (both EDA and targeted)
- 9.2 Test and debug scripts/queries on sample data
- 9.3 Run the developed scripts/queries on the full dataset
- 10.0 Perform Targeted/Focused Analysis
- 10.1 Apply specific analytical techniques based on question types and initial EDA findings
- 10.2 Perform statistical tests or implement models as appropriate
- 10.3 Conduct deeper investigations into areas of interest identified in initial exploration
- 10.4 Iterate on analyses as needed, refining approaches based on interim results
- 11.0 Interpret & Analyze Results
- 11.1 Review outputs for reasonableness and consistency
- 11.2 Interpret results in the context of audit objectives, questions and hypotheses
- 11.3 Validate findings through cross-verification and subject matter expert consultation and auditee
- 11.4 Identify areas requiring further investigation or additional data
- 11.5 Synthesize insights from both EDA and targeted analyses
- 12.0 Documentation and Iteration
- 12.1 Document insights, methodologies, and decision points throughout the analysis process
- 12.2 Maintain a log of all analyses performed, including unsuccessful attempts and rationale
- 12.3 Prepare preliminary findings for review with key stakeholders and auditees
Key Deliverables
- Documented scripts, for example Python code or Jupyter Notebooks, and logs (for both EDA and targeted analyses)
- Detailed analysis of data sets and outputs
- Information visualization (charts, graphs, tables) illustrating key findings
- Preliminary findings report, including:
- Summary of EDA results
- Outcomes of targeted analyses
- Identified patterns, anomalies, and areas of interest
- Analysis methodology documentation, including decision points and rationale