Pre-conference event: May 13

09:00

The journey to profitable digital trading models: Exploring the data infrastructure and methodology needed to rapidly train, test, and deploy proprietary digital trading models

• Cloud based data storage and management optimized for trading signal development
• High performance compute infrastructure to support model development, training and deployment
• Geospatial and alternative data sources and modalities
• Developing digital supply and demand factor models 101
• Identifying a major supply disruption prior to market knowledge of the event
• Bespoke weather forecasting techniques
• Forecasting rig completions and production in an individual basin 
• Forecasting a future charter rate for an ocean vessel class
• Deep dive: Forecasting gasoline import demand using multimodal data fusion and machine learning

James Orsulak, Senior Director of Business and Sales, DESCARTES LABS 

10:30

MORNING BREAK

11:00

Machine learning in power trading including application of machine learning and neural networks in load forecasting and scheduling

• Using fuzzy classifier to manage pipeline optimization and electricity network analysis (allow identification of opportunities in congestion ,capacity trading etc)
• Using machine learning for renewables systems analysis.
• Using genetic algorithms   for logistics management in physical energy asset (electrical networks , gas ,coal ) using practical examples to illustrate and illuminate how to build genetic algorithms for the above use cases
• Managing complex data sets and big data: managing standard energy assets data sets such as networks, contracts , asset price curves , drilling info , logistics and supply chain data  in conjunction with very large scale data
• Introducing graph , array oriented , gpu accclerated , object ,embedded  and other times of novell databases technologies ( including Hadoop and hpc/Hadoop hybrids ) to address the above big data and complex data problems

Sid Dash, Research Director, CHARTIS RESEARCH

1:30

Visualizing energy prices with R: From raw data to an interactive dashboard of modeled  relationships

• The role and practice of data visualization in energy price research, using tools and packages in the R programming language
• Starting with raw price data, exploring  model and visualizing it, and finally creating an interactive dashboard to house that work in a compelling format
• Different parts of the R toolchain and how they fit into the broader data science paradigm
• Live coding examples

Jonathan Regenstein, Director of Financial Services, RStudio

3:00

AFTERNOON BREAK

3:30

Using data science and machine learning for energy investments and risks analysis

• What is the existing approach?
• What are the pros and cons?
• How are the outcomes like?
• Do we back-test? If yes, how does it compare to the inception?
  - What could be wish list items?
  - How do we identify data sources & collect data?
  - What modeling approaches are considered?
  - How could we leverage data science & machine learning (Artificial Intelligence)?
• Discuss a practical example and future possibilities

Akshay Singh, Manager, Quantitative Analysis, WASHINGTON GAS

HARVEST SESSION & Q&A

5.10

End of the workshop.