Junior Data Scientist
Move Loot, Inc
- Data continuity and integrity
- Refactor and maintain Redshift schemas to accommodate an increasingly complex set of back end applications, a rapidly growing team and constantly changing business priorities.
- Training and Support
- Hosted Data Office Hours to promote independent data exploration
- These sessions provided interpretation and QA for self-pulled reports, and supported business users in obtaining the data they need.
- Hosted and recorded a biweekly series of workshops and analytics demonstrations.
- These served not only as a means to proliferate our methods for data analysis and improve transparency of our analytics, but also as onboarding tools for future team members, analysts and data scientists in particular.
- Onboarded all new team members in our analytics stack, data sources, and training in all analytics tools and processes.
- Full stack dashboard Flask application
- Set up an end-point and a redis cache to catch entities sent from production in realtime and render in charts.
- Used dc.js, a library build on d3 to quickly render custom charts populated by a three day spotlight of our transactional database.
- Product recommendation algorithm
- Permanently improved average cart size, inventory sell-through, and conversion rates with a SQL powered algo, which periodically pushed product ID's to a redis cache with an http endpoint to rapidly serve product recommendations to users browsing the site.
- Marketing + Product Analysis
- Reduced CAC with feature analysis using grid search to identify key indicators of successful marketing strategy.
- Collaborated with web team to A/B test user flow; reduced bounce rates and grew the top of the funnel.
- Bayesian statistical analysis of referrals program, grew referrals by 25% in a week by A/B testing CTA placement and referral bonus
- ROI report of offline direct mailer campaign, using Levenshtein distances for address matching.