AnalyticsData Science
Turning Conversation Data Into Product Recommendations
Analyzing CAISY beta interactions and learner cohorts to understand user behavior, recurring queries, and opportunities for better prompting and audience targeting.
- Organization
- Skillsoft Corporation
- Timeline
- July 2024 – September 2024
- Location
- Remote
- My role
- Product & Content Strategy Intern — conversation analytics and content strategy.
The problem
CAISY beta produced conversation logs and learner-engagement data; the team needed to understand real usage and implications for prompts and targeting.
Why it matters
Product decisions improve when grounded in real user behavior.
Data & inputs
- CAISY beta conversation logs.
- Learner cohort and engagement data.
- Percipio metadata and technical learning paths.
Approach
Analyzed logs and cohorts in Tableau and Power BI; translated patterns into prompt and audience-targeting recommendations; curated Percipio metadata.
- 01Conversation logsCAISY beta + cohorts
- 02AnalysisRecurring queries, engagement
- 03VisualizationTableau / Power BI
- 04RecommendationsPrompt + targeting guidance
My contribution
- Analyzed beta conversation logs and learner cohorts.
- Visualized engagement patterns and recurring queries.
- Translated findings into prompt and audience-targeting recommendations.
- Curated Percipio metadata and technical learning paths.
- Developed a Systems Engineering benchmark planbook for Instructional Design.
Technical details
- Tableau, Power BI, Excel for analysis and metadata curation.
- Cohort comparison across role-play scenario responses.
Challenges & decisions
- Conversation logs are unstructured; judgment is required to find product-relevant patterns.
Results
- Delivered prompt and audience-targeting recommendations.
- Produced Systems Engineering benchmark planbook for Instructional Design.
Limitations
Findings reflect the CAISY beta period. Analytical inputs to product decisions rather than measured production outcomes claimed here.
What I learned
Connecting data analysis to product thinking — recommendations an audience can act on.