Data Science Unit

Corporate Services – Hedging Programs

Brickell Analytics builds models to forecast the direction of global macro markets: equity indices, currencies, oil, and rates. The objective of the models is to assist interested parties with their capital deployment.

These initiatives can also be applied to corporate hedging programs. For example, the models could be used to hedge the direction of interest rates, currencies utilized, and oil at predetermined periods.

Our primary focus is on the following assets:

Euro – EURUSD, Japanese Yen – USDJPY, Swiss Franc – USDCHF, Canadian Dollar – USDCAD, Australian Dollar – USDAUD, West Texas Intermediate – OIL, Gold – GOLD, Current Spot Copper – COPPER, Treasury Curve – (2YR,5YR,10YR), and the S&P 500 Index.

Corporate Consulting Services Overview

We began our services in 2011 primarily focused on core mathematics paired with price patterns and crowd psychology analysis. Over time, the needs of our clients in the asset management industry have grown along with the rise of quantitative model trading strategies. Our services have evolved along with them, so that our customers can make educated decisions with their assets.

Our forecasting methods can assist multinational companies with hedging foreign exchange volatility and with input budgeting. We spot patterns in the macro markets that are difficult to discern with conventional research and fundamental analysis. These unbiased models could help you complement your current fundamental or corporate views on the economy. When carefully planned, these patterns might complement a firms’ decision making process and financial modeling analysis.

In essence, we pair big data with models to help clients solve problems in different industries. We are capable of tailoring our tools and deployment paths to address unique needs across sectors.

  1. Pharmaceuticals & Healthcare that have multi-currency exposure and international presence. We could provide alternative data to help predict how input variables could affect the bottom line.
  2. Asset Management: add a layer of unbiased analytical processing to assist the investment decision-making process. Past projects implementations include the use of the following data sources: Federal Reserve Data, FDIC Public Data, Bloomberg, Quandl, Industry Databases, Social Media, and unstructured data (NLP for sentiment analysis).
  3. Real estate: Traditional valuation methods focus only on linear relationships. We help you design models with multivariate inputs based on non-linear variables. Today’s dataset comes in all formats, from traditional surveys to mobile phone activity that can help us form models that challenge conventional thinking. Do you have a program in place that monitors the social media presence of your tenants that could help you detect issues ahead of time?
  4. Manufacturing: we can work with all manufacturers in the areas of production, logistics, and task optimization.
  5. Hospitality: Assist mid-size hotel companies looking to create revenue and utilization models to better optimize pricing.
  6. Automobile Industry (including parts): companies in the automobile manufacture chain can take advantage of our modeling capabilities. We are able to model rates, consumer sentiment, economic indicators, foreign exchange, and unstructured data to help in the decision-making process. Create checklists, such as, economic indicators, forex volatility, and research forecasts. Additionally, we can tailor models used across supply chains to help clients set structures on pricing, costing methods, inventory levels, and after market parts service levels.

Background on Data Science Team

We have experience working with business process players to design models that actually resemble your company needs. The objective is to work with various business lines to identity needs and/or issues, collecting datasets to then build models with advanced analytics techniques.

Technologies

  1. Python
    One of the leading languages for data analytics; we have over 10 years of expertise in the use of Python.
  2. R
    Mostly used in academia, we have the capabilities to implement a project in R.
  3. Machine Learning
    We are familiar with many of the libraries out in the marketplace. Some of the projects are based on TF(python only), Scikit-learn, and Keras.
  4. Notebooks
    We love Jupyter notebooks! If your project requires the use of Zeppelin, Spark, or Databricks we are also ready to tackle it.
  5. Other
    We have knowledge in Java, Spark, SQL, and in the several widely recognized libraries used in data science. We can also work on Windows, MacOS, and Linux environments in the cloud or on premise.

Focus

As data-science practitioners, we can assist with data engineering challenges. Some of the services that can be implemented are as follows:

  1. Analytics and Visualization: turn data into insight via visualization models that can help your organization make decisions in a shorter time frame.
  2. Statistical Modeling: help with the design of a data gathering process that pairs well with some of today’s most efficient machine learning solutions.
  3. Education: we have done lectures at private and public venues, and are adept at educating in a group setting.
  4. Staff augmentation: interested in having someone more full time? Our relationships with some of the best schools and students in statistics and math has allowed us to tap data scientists across the world

Data Science Services

Here are some of the projects that we can implement via our services:

Data Science Specific:

Causal inference and experiment design:

Through in-market experiments or carefully designed testing on historical data, determine what factors have the strongest influence on your bottom line. Applications:

  1. measured impact of marketing campaigns
  2. ordinal regression models for employee skill growth on human capital development initiatives

Forecasting

We build prediction models to help you make better business decisions.

Applications:

  1. Optimize inventory and sales through accurate demand forecasting
  2. Revenue management
  3. Asset predictive maintenance
  4. Time series analytical models

Classification

Categorize data into useful segments to allow differentiation Applications:

Sensitivity Analysis

Quantify the uncertainty surrounding your important business decisions to better assess risk.
Applications: calculate the risk of costly stock outs or a product not selling as well as planned

Optimization

Use linear or non-linear optimization algorithms to maximize revenue and/or minimize costs
Applications: calculate the price to maximize sales throughout a time period, minimize supply chain costs

Simulations

Develop deep understanding of the possible outcomes from a crucial business choice. Bayesian statistics and Monte Carlo Markov Chain methods to create thousands of simulations of likely outcomes to identify course of action with associated level of risk.

Data Engineering

  1. Convert messy, unstructured data into structured data from which you can extract useful insights
  2. Save hundreds of hours of manual labor by automating data collection, storage, and processing tasks.

Data Visualizations

Provide decision makers and analysts with interactive data. Tools implementation via PowerBI, Tableau, or custom JavaScript based dashboards.
Schedule automated reports in place of time consuming ad hoc queries.