Luke Taylor Email & Phone Number

Currently working at Companies House as a Data Scientist within the Data Science team. Learning AWS, Power BI, and Alteryx, in addition to growing my analytical, forecasting, and AI knowledge and ability.Postgraduate studies allowed me to delve deeper into my love of data and to develop creative solutions to complex problems. The course entailed many projects using real-world data, statistics, modelling, machine learning, and AI, for example:• Stock Price Prediction – Implementing LSTM and GRU recurrent neural networks for a one and seven day forecast of the S&P500 index in Python using sklearn and keras libraries and applying data pipelining.• Investigation of shots taken in the NBA 2014/15 season – Performing parametric and non-parametric tests to draw conclusions on player performances, OLS Linear Regression modelling defender distance with shot distance, Logistic Regression determining success with parameters chosen via variable selection all in R.• Mechanic Garage Simulation in “Simul8” – Simulation based on demand distributions. Recommendations made based on resource utilisation.• Weekly Forecast of Jakarta Ambulance Service – A comparison of forecasting methods to provide a 7-day forecast via the ARIMA method.• Supply Chain Modelling of fresh vegetable production – Using Excel’s OpenSolver add-in I formulated a stochastic linear programming solution to minimised the cost of meeting customer demand based on uncertain farm supply and customer demands.• Revenue Prediction via Machine Learning methods – Decision Tree, Nonlinear SVMs, and K-Nearest Neighbours classification in parallel with Random Forest and Grid Search were compared based on AUC, F-measure, and Recall performance metrics.• Optimisation of Tidal Range Scheme Energy Generation – My dissertation involved producing a simulation of the proposed West Somerset Lagoon, a control function for operation, and creating a Genetic Algorithm to maximise efficiency of energy generation in Python. Then producing a report of the study with a thorough literature review. The algorithm succeeded, improving the energy output by 13.6% from the naïve baseline solution.For more information on any of these projects please feel free to contact me.I enjoyed my time at university and the opportunities it presented to develop interpersonal, academic, and professional skills before entering the world of work soon after.