Data and Market Research Specialist

profile

With almost 10 years of experience in financial writing and market research, I leverage data analytics and research skills to communicate a deeper understanding of company performance across a range of industries, including the consumer goods, hospitality, financial services and insurance sectors.

I most frequently use Python, R and SQL for data analysis purposes. Specifically, I use data visualisation and statistical techniques to better understand consumer and market trends within a specific industry, such as the impact of price changes on consumer demand, or the effects of seasonality on revenue growth.

Examples include:

Analysis of hotel revenues: Implemented data manipulation and visualisation techniques to illustrate quarterly RevPAR (revenue per available room) fluctuations and total room revenue by hotel brand. Leveraged data analytics to identify hotel brands with highest revenue growth across the luxury-priced segment.
Lifetime Value (LTV) of postpaid telecommunication customers: Used average revenue per user and churn rates to calculate lifetime value of customers by quarter across the postpaid segment of major telecommunication companies. Analysed seasonal fluctuations to identify financial quarters with the highest LTV over a multi-year period.
Loss ratio analysis of insurance companies: Leveraged probability methods to generate loss ratio simulations for insurance companies specialising in the Property & Casualty sector. Provided assessment of risk based on the impact of continued inflationary pressures and extreme weather events on the demand for property insurance.

In addition to my work as a financial writer, I have also used data science methodologies to deliver effective business intelligence solutions to companies across a range of industries. Methodologies that I have used in providing solutions to clients include:

Probability and Risk Analysis
Supervised and Unsupervised Machine Learning Techniques
Time Series Analysis

Samples

Forecasting Hotel Revenue: Predicting ADR Fluctuations with ARIMA

Average daily rates represent the average rate per day paid by a staying customer at a hotel. This is an important metric for a hotel, as it represents the overall profitability of each customer. In this example, let us see how average daily rates can be forecasted using an ARIMA model.

Handling Imbalanced Classification Data: Predicting Hotel Cancellations

When attempting to build a classification algorithm, one must often contend with the issue of an unbalanced dataset. An unbalanced dataset is one where there is an unequal sample size between classes, which induces significant bias into the predictions of the classifier in question. Let us see how a Support Vector machine can be used to classify hotel booking customers in terms of cancellation risk.

Hotel Analytics: Calculating Revenue, RevPAR and GOPPAR Using SQL

The hotel industry relies on several unique KPIs to gauge performance. One of the most important of these is RevPAR - which stands for revenue per available room.

In addition, hoteliers can also use GOPPAR (or Gross Operating Profit per Available Room) to calculate room profitability. Often, hotel businesses can find it challenging to keep up with these metrics and identify key revenue and profitability trends over time. However, using SQL and data visualization together can be quite an effective way of analysing these metrics.

Shiny Web App: Analysing hotel metrics using pyplot

This is a Shiny Web App developed using a bubble chart generated using pyplot to analyse metrics including RevPAR, ADR, occupancy and number of rooms using publicly available Q3 2023 IHG earnings data.

Technical Skills

Cloud: AWS, Azure

Languages: Python, R, SQL

Machine learning libraries: InterpretML, PyMC3, scikit-learn, statsmodels, TensorFlow

Platforms and relevant tools: PyCharm, Jupyter Notebook, pgAdmin4, RStudio, Git, Docker, Linux

Visualization libraries: ggplot2, matplotlib, seaborn

Training Courses

Time Series Forecasting with Bayesian Modeling. LiveProject series produced for Manning Publications.

Devised liveProject series to illustrate modelling of time series shocks with Bayesian Dynamic Linear Modeling, modeling of posterior distributions with PyMC3, MCMC sampling with Prophet, and Structural Time Series Modeling with TensorFlow Probability.

TensorFlow 2.0 Essentials: What’s New. Video seminar produced for O’Reilly Media.

Illustrated use of eager execution and AutoGraph, as well as use of tf.keras for neural network modelling across classification, regression, and time series datasets.

Business Analytics with R — Statistics and Machine Learning. Video series produced for O’Reilly Media.

Illustrated use of data manipulation techniques, regression analysis and hypothesis testing, along with classification and regression-based machine learning techniques.

Michael Grogan