1. Introduction:

Our goal is for the intern to have an educationally stimulating, professionally rewarding, and personally fulfilling experience while interning at Vistex. The opportunity to work with interns can have far-reaching implications for future Vistex product development. We anticipate this collaborative venture to benefit both parties.

2. Outline:

The Data Science Team at Vistex is interested in including predictive index-based pricing models for the petroleum industry in its pricing setting solution for future offerings. Therefore, we would like the intern to research what index-based pricing is and how index-based pricing is being defined, used, and modeled in the petroleum industry. The intern will collect the needed datasets from public sources or simulate some fields if needed. The intern will set up a base model to include needed features / variables and determine the best model for forecasting petroleum costs. Further model improvement and comparison will be highly encouraged. The intern will conclude their work with a PowerPoint presentation on the overall process, obtained findings and results, and any further insights or future improvements.

• Plan:

1. • WEEK 1 – Orientation and Onboarding
     • Meet with HR.
     • Meet with mentor.
     • Understand Vistex’s mission, vision, and values.
     • Become familiar with the Data Science Team and their work scope.

1. • PROJECT SCOPE – This is the bulk of the 10-week program. The timeline is provided as a suggestion but is somewhat flexible to accommodate the interest of the interns.

1. • • In-depth research into the index-based pricing (1-2 weeks)
       • Identify how index-based pricing is defined and calculated.
       • Identify how index-based pricing is used & applied in the petroleum industry.
       • Investigate what datasets and fields need to be collected for modeling.
       • Try to become the resident expert in index-based pricing.

1. • • Dataset collection & fields simulation if needed (1 week – may need refresh every 2-weeks)
       • Collect needed datasets and fields from public sources.
       • If some fields are unavailable but important, try to simulate these fields.

1. • • Set up a base model for forecasting the price index / indexes (1-2 weeks)
       • Identify needed features / variables for the forecast model.
       • Feature engineering is recommended to derive more features.
       • Set up a base model on forecasting petroleum price index.
       • Determine the adequate equation on forecasting petroleum costs from the forecasted petroleum price index.

1. • • Model improvement & Comparison (1 week)
       • What other models can be introduced?
       • What is the model improvement compared to the base model?

1. • • Derive final price-index model (equation) that can track index changes (1-2 weeks)
       • What are the tracking indexes with changes along time?
       • How do those changes reflect in the final price?

• • • Prepare & present final presentation of index-based pricing research, findings and results, and insights gained (1 week)
      • On-site presentation to the Data Science Team and other interns.
      • Summarize how Vistex could implement the index-based pricing in the future.

3. Misc:

• Quick twice-weekly check-ins, or as needed, with their mentor for support and guidance.
• Expand the intern’s professional network of colleagues in the office.
• Mentors from Vistex that will guide you in the project.

4. Useful resources

• Introduction to index-based pricing: https://www.coursera.org/lecture/uva-darden-bcg-pricing-strategy-cost-economics/index-based-pricing-FieKW
• Introduction to crude oil: https://www.cmegroup.com/education/courses/introduction-to-crude-oil.html

5. Dataset candidates

• Public data: Collected from US Bureau of Labor Statistics, PPI
• Forecast Data: US energy information administration, EIA
• Market Data (WTI): Yahoo Finance Year (2000-2023), weekly data
• Market Data (RBOB): Yahoo Finance Year (2000-2023), weekly data
• Chicago City (regional gas stations (shell)/(BP) prices for one year)
• •  

6. Eligibility Criteria: 

• Student must be enrolled in the MS in Data Science degree and have completed all the foundation courses
• The students must be available to work for 10 weeks during Spring quarter for 10 hours per week
• Students must have a minimum GPA of 3.0.

7. Application Requirements: 

• Informal transcript
• Recent class project student worked on
• A data science project done outside of class (if applicable)
• A 3-minute video recording of you presenting your favorite data science project
• More details can be found in the application link below. 

8. Apply Here:

• Deadline to apply: March 27, 2023. 

• Please apply using this link: https://tinyurl.com/4cptvm4

• For any questions please email: iustun@depaul.edu

The first project was titled “Agent-based modeling of price guidance”. A team of three data science students worked on the data science project under the supervision of Prof. Ilyas Ustun. This project can be used to fulfill the capstone requirement for the MS in Data Science degree program.

Price guidance is a common concept used in Business-to-Business pricing. It consists of a couple of reference prices like a floor price, a target price, and an expert price that are used to provide guidance to the sales agent who is trying to respond to a request by a customer for a product or service. The reference prices are calculated based on the outcomes of previous quotes for similar products to similar customers.

Imagine that this quoting process is automated. The customer asks for a product, then an algorithm representing the supplier responds with a quote and finally the customer either rejects or accepts, and that’s the end of the quoting process. What would be a good algorithm for the supplier? How to assess using repeated simulations of this process how this algorithm performs? 

Vistex designs a repeated transaction with one supplier and a series of customers with the same willingness-to-pay distribution. The supplier does not know the willingness-to-pay distribution. The supplier can make one price quote per request and based on that the customer will either accept or reject. Vistex provides a simple baseline model for the supplier as a reference point. Possible tasks for the intern:

(i) Determine optimal estimation and optimization method

(ii) Determine impact of competition, and let suppliers compete with different methods

(iii) Determine impact of segmentation, and assess value of cross segment information

The students must be available to work for 10 weeks during Winter quarter for 10 hours per week at a pay of $16 per hour. Start date will be January 2, 2023.

1. Preferred Knowledge, Skills and Abilities:
   • Strong analytical skills
   • Experience with Python programming
   • Experience with statistical analysis and machine learning techniques
   • Experience with optimization methods is a plus
   • Interest in business processes and go-to-market strategies is a plus
2. Eligibility Criteria:
   • Student must be enrolled in the MS in Data Science degree and have completed all the foundation courses
   • The students must be available to work for 10 weeks during Winter quarter for 10 hours per week
   • Students must have a minimum GPA of 3.0.
3. Application Requirements:
   • Informal transcript
   • Recent class project student worked on
   • A data science project done outside of class (if applicable)
   • A 3-minute video recording of you presenting your favorite data science project
   • More details can be found in the application link below.