Prediction of Off-Shore Rig Utilization Rate

Developers

• Han Yang
• John Hawkins
• Shilpa Muralidhar
• Michael Harper

Background

The price of contracting a drilling rig is highly affected by the rig utilization rate. An offshore drilling rig company is considering making a bid on a tender that has been received. The rig is currently in a ‘warm stacked’ mode and costs $15,000 per day to operate. In order to put it back to work, an additional $80 million is needed plus an operational cost of $150,000 per day. What day rate should be charged for a 4-year term to avoid underbidding the market?

Objectives

1. Identify the potential factors impacting global off-shore rig utilization rate
2. Utilize machine learning algorithm to establish a model to predict global off-shore rig utilization rate.
3. Validate the model by comparing prediction vs actual.

Tools

• Python
• Tableau

Results

1. Feature Importance (SHAP Values)
2. Xgboost Model
3. Long Short-term Memory (LSTM) Model

Conclusion

1. The most influential factor is world oil consumption followed by active rig count and natural gas price.
2. Although both Xgboost and LSTM models did a good job on predicting, LSTM model is preferred to solve the problem, since it could predict for the next month.

Tableau Public Website

https://tabsoft.co/366KwAm