Insurance, Data Science, Python and more

Chrysalis Insurance

Reading Time 1 mins

Chrysalis is a consortium of energy insurers with a mission to develop commercially sustainable insurance products relevant to companies in the energy industry.

It is supported by a joint venture development company through which the insurers collaborate to design insurance policies.

This is a company I have been working with the past 18 months - the team recently launched their first new product called “Operability”.

Chrysalis Insurance Ltd

@ James Poynter Tue, Aug 4, 2020

Insurance Innovation 2020

Insurance

Reading Time 1 mins

I will be speaking at the Insurance Innovator summit in November

Insurance Innovators Summit 2019

@ James Poynter Sun, Aug 11, 2019

GeoInsurance Europe 2019

Insurance

Reading Time 1 mins

I will be discussing data analytics and modelling at the upcoming GeoInsurance Europe conference in January.

GeoInsurance Europe 2019

@ James Poynter Fri, Nov 23, 2018

Conway's Game of Life with Matplotlib, Scipy, and numpy

Python

Reading Time 5 mins

In this post we will develop a Python implementation Conway’s Game of Life, set in a donut shaped universe!

The post will utilise numpy, matplotlib’s animation features, and Scipy’s 2D convolution tool kit. It also provides a nice demonstration of what can be achieved with just a few lines of Python!

There are a series of 5 longer more colourful video examples at the end of the post

@ James Poynter Sat, Oct 20, 2018

Time-Series with Gradient Boosted Models London Fire Brigade Call-outs

This post explores the application of gradient boosted algorithms to time-series modelling, as we try to predict the number of London Fire Brigade Call outs using seasonal patterns, and 3rd party weather data.

@ James Poynter Sat, Oct 6, 2018

Time-Series with Gradient Boosted Models London Fire Brigade Call-outs

Multi-Step Grid-Search Pipelines: A Lazy way to tune XGboost?

XGBoost

Python

Machine-Learning

Data Science

Reading Time 6 mins

A couple of years ago I read a blog post on Analytics Vidhya Complete Guide to Parameter Tuning in XGboost (with codes in Python). The original post uses a multi-step grid-search to tune an XGBoost model. This post will develop a simple “pipeline tool” to automate this sort of tuning.

Knowledge of, and interest in machine learning is beneficial to follow along

@ James Poynter Sat, May 26, 2018

Building Static Websites with the JAMStack

JAMStack

Software

Reading Time 6 mins

JPytr.com has recently been rebuilt using HUGO and the JAMStack. This post discusses gives a brief overview and introduction to the JAMStack.

Familiarity with version control and the basics of web-development will be helpful to follow along.

@ James Poynter Sat, May 26, 2018

Building Static Websites with the JAMStack

Domain Knowledge for Data Scientists

Domain-Knowledge

Data-Science

Insurance

Reading Time 7 mins

Data science is often described as the overlap of domain knowledge, computer science, mathematics and statistics.

@ James Poynter Mon, May 7, 2018

Analysing Geographic Data with Folium

Folium

Geographic-Data

Data-Science

Reading Time 211 mins

As far as analytic techniques go, visualising geographic information using maps has endured the test of time. Map making was once the art of the skilled cartographer, however modern technologies mean that creating rich interactive visualisations in only a few lines of code away.

In this post we will develop a interactive mapping tool using the Folium Python library to assist in analysing geographic risk.

@ James Poynter Mon, May 7, 2018

Introduction to Specialist Insurance for Data Scientists

Specialist Insurance

Insurance

Reading Time 8 mins

Insurance is about risk management and risk transfer. The basic concept is simple, a client (the assured) enters into a contractual arrangement with an insurer, who agrees that in exchange for an upfront cost (the premium), that they will restore the financial position of the assured (indemnify) in the event that they sustain a loss that is covered under the terms of the contractual arrangement (a claim).

@ James Poynter Mon, May 7, 2018

Predicting Cycle Hires with Facebook Prophet

Facebook Prophet

Time-Series

Open-Data

Data-Science

Reading Time 5 mins

In this post we will be using Facebook’s Prophet time series analysis api to forecast daily cycle rentals in London.

@ James Poynter Sun, May 6, 2018

Predicting Cycle Hires with Facebook Prophet

Guide to Random Forests: see the wood for the trees with LIME and Jacknife variance

JackKnife Variance

Random Forests

Data-Science

Reading Time 24 mins

This tutorial will provide an introductory guide to tree based models, and random forests. In the second half of the tutorial we will look at how to explain models using Local Interpretable Model-Agnostic Explanations (LIME), and how to calculate confidence intervals around random forest predictions, using an approach called infinitesimal jack knife variance.

@ James Poynter Sun, Apr 29, 2018

Data Science Festival Presentation

Presentation

Data Science

Insurance

Reading Time 1 mins

Presentation at Data Science Festival in London 2017.

The London Insurance Market is the world’s leading hub for specialist insurance, a place where underwriters and brokers make markets for and trade new, complex, and large risks. This talk will give: an overview of the multi-billion dollar market, discuss the complexity of offshore energy risk, and the exciting potential applications of data science techniques within the specialist insurance sector.

Data Science Festival

@ James Poynter Fri, Jan 20, 2017