data science projects to date
An overview of the data science projects I have completed both inside and outside of work.
My projects so far include:
Statistical modelling
- Fraud model for Zopa - designed model to detect fraudulent applications.
- Propensity modelling Zopa - designed model to predict certain customer behaviours.
- Regression algorithm - coded vectorized cost function, gradient descent function and normalized equation function for regression and classification predictions in Octave.
- Neural network algorithm - coded feedforward, regularized cost function and backpropogation for classification predictions in Octave.
- Predictive modelling - created a survival prediction model for Titanic passengers in R.
- Twitter tweet analysis - deduced the happiest U.S. State by establishing the sentiment and location of Twitter tweets in Python
Analytics
- Conversion optimisation with Splunk for Zopa - combined Splunk and SQL data to aggregate end-to-end user journeys per device type.
- Email campaign Zopa - campaign design, implementation and success analysis.
- Remarketing insights for AnyRoad - designed and implemented remarketing strategies using customer segmentation in R, and Google Analytics.
- AB testing for AnyRoad - Optimized page features using Google Analytics.
Data visualisation
- Company health check dashboard Zopa - shows key company metrics, built in Flask and generates daily email.
- Team progress visualisation at Zopa - live view of progress per team member, built in Flask and fed via SQL.
Data infrastructure
- Data landscape map at Zopa - mapped the different data themes of the company.
- MapReduce - adopted Map and Reduce functions to invert indices, create relational joins and perform social network analysis in Python.
- Relational database design - designed project database structure using class diagrams, relation schemes and normal forms.
Opportunities & Collaboration
Always feel free to get in touch with opportunities.