David Rhode Email & Phone Number

London Area, United Kingdom

Stella McCartney is an international fashion brand known for its strict adherence to ethical sourcing and manufacture. A year after being hired, I was promoted to Senior Data Scientist. I was originally brought in to help grow the ecommerce side of the business, but my role has now expanded to also support the Head of Retail (in charge of nearly seventy stores globally).• Forecast Customer Lifetime Value in a non-contractual situation using the Python Lifetimes package. Used the insights derived from CLV calculations to assess the risk/reward of introducing a new range of entry-level ‘gateway’ products, with the aim of inducing new customers to graduate to purchase ‘high value’ items.• Used Google’s LightweightMMM package to generate sales forecasts. Weekly forecasts had MAPE ~8%, monthly forecasts had MAPE < 5%.• Used Geopandas to analyse ecommerce sales based on customer zip codes. This was combined with demographic data and luxury ‘door counts’ to support the decision around where to open a new US store.

London, England, United Kingdom

I was hired to help this iconic fashion brand to get more value from its data, and to begin the transition to becoming a genuine ‘digital-first’ business.• Used Bayesian statistics (PyMC package) to model the return rates of ‘Ready-To-Wear’ items. This model was used to search for a ‘switch-point’ that may have coincided with the introduction of new ‘anti-wardrobing’ tags in ecommerce shipments.• Helped the CRM manager to create new customer segments for targeted communications using Basket Analysis and Customer Churn Modelling.• Used a HuggingFace Transformers model to perform Sentiment Analysis on scraped Twitter traffic to identify key online influencers.• Liaised with parent company LVMH to initiate a move towards best practices for data storage and access.

London

EMERGEiQ delivers game-changing insights by offering data science and AI as a service to small and medium sized enterprises. I was hired to work on an app for a client in the Health/Wellness sector, with the aim of serving each user with the most relevant content for their needs. I then picked up a legal project, for a client who wanted to use NLP to extract insights from court documents. - Trained a classifier using the HuggingFace Transformers library to determine whether a user might be suffering from depression based on their journal entries.- Used DoWhy and EconML libraries to model the effect of promotions for a retail client.- Built a Recommender System to serve audio and video content to app users.- Built an anomaly detector using an Encoder/Decoder structure in Keras/TensorFlow, to identify interrupted sleep patterns in wearers of Garmin watches.- Used the Pygatt library to source streaming data from Muse headbands (Bluetooth Low Energy).- Used a Hidden Markov Model from TensorFlow Probability to simulate the changing internal mood states of an app user over time.- Trained a multi-label classifier to determine the relevant claims and verdicts from the Judgments and Reasons sections of historic legal texts.- Built a search engine using a Transformer model, allowing app users to actively find useful content from a library of audio and video.- Worked closely with engineering team to transition project from Proof-Of-Concept to Minimum Viable Product stage.- Mentored summer interns.- Developed client dashboard in Google Looker Studio.

A twelve week full-time course, covering the key aspects of supervised and unsupervised machine learning with Python. My five-week capstone project used geographical features to predict London property prices, and allowed me to conclude exactly how much of a property’s value can be ascribed to its location as opposed to features such as size and quality. Geographical data was handled with GeoPandas, key models included Linear Regression, Random Forest, XGBoost. This process allowed me to ascribe relative values to all 200k London postcodes (including the 130k that didn’t appear in the original data), and to write a function to ‘move’ a property from one postcode to another, and calculate its relative change in value. In the future, these geographic features could be layered on top of more traditional data to allow estate agents to conduct more realistic evaluations of a property’s value in advance of a sale.Please see below for the project README.

With a friend from university, I co-founded the UK’s leading ethical fine jeweller. Ingle & Rhode specialised in engagement rings and wedding bands, made to a premium standard, from traceable, ethically-sourced materials. The company operated a ‘bricks and clicks’ business model. This allowed a level of customer service that isn’t possible with pure ecommerce businesses, but avoided the big overheads associated with traditional retail.- 66% conversion rate from appointments to orders- Raised over £0.5m through equity crowdfunding- Grew turnover past £1.6m- 5-star rating on Trustpilot- Cited as an authority on CSR by The Guardian, Professional Jeweller, The Independent, Retail Jeweller- Launched Fairtrade Gold Design Awards in conjunction with the Goldsmiths Council- Finalist, UK Jewellery Awards 2015, 2016

United Kingdom

My responsibilities included directing dramatic reconstructions; conducting interviews with contributors; storyboarding and commissioning special effects; writing scripts; directing editors; negotiating access to stories; writing development documents; pitching to commissioning editors; and managing the logistics of foreign shoots.DARLOW SMITHSON August 2006-December 2006‘Seconds From Disaster’, (1 x 1hr) for the National Geographic Channel.A forensic look at the events that led up to a huge explosion at BP’s Texas City refinery in 2005, which killed fifteen people. FOLIO/MENTORN February 2006-June 2006‘ASBO Fever’, (1 x 1hr) for Sky One.A light-hearted examination of some of the most unusual ASBOs to have been handed out. REDBACK PRODUCTIONS August 2005-December 2005Conspiracies, (1 x 1hr) for Sky One. The story of the disasters that hit the Kennedy family, featuring the deaths of JFK and Bobby, and Teddy’s disgrace at Chappaquiddick. DIVERSE PRODUCTIONS February 2005-June 2005The Trouble With Young People (1 x 1hr) for the BBC. A polemic fronted by journalist Carole Malone, on the rise of liberal parenting. FLASHBACK TELEVISION January 2004-November 2004The Porn King Vs the President (1 x 1hr) for the BBC. Rob Black, a San Fernando pornographer, is due to stand trial. His case highlights the attempts by evangelical Christians to blur the boundary between church and state.

Bachelor Of Arts - Ba, History

Data Science Immersive