Vincent Chung

Vincent Chung

Austin, Texas Metropolitan Area
7K followers 500+ connections

About

Currently a solutions engineer helping leaders and organizations drive decisions and…

Articles by Vincent

Activity

Join now to see all activity

Experience

Education

Volunteer Experience

  • University of Notre Dame Graphic

    Capstone Sponsor, MS Business Analytics class of 2021

    University of Notre Dame

    - 4 months

    Education

    Provided capstone project sponsorship to graduate students of the Master of Science in Business Analytics program, Class of 2021.

  • Georgia Institute of Technology Graphic

    Capstone Mentor, MS Analytics class of 2019

    Georgia Institute of Technology

    - 4 months

    Education

    Provided capstone guidance and held office hours to graduate students in the Master of Science, Analytics program during the capstone phase of the 2019 program.

  • MATTER Graphic

    Workshop Instructor

    MATTER

    - 7 months

    Science and Technology

    Designed and led workshops around fundamental machine learning concepts and its applications within the healthcare space.

  • Illinois Shotokan Karate Clubs Graphic

    Tournament Referee

    Illinois Shotokan Karate Clubs

    - 6 years

  • Ronald McDonald House Charities of Chicagoland & Northwest Indiana Graphic

    Volunteer Prep Cook

    Ronald McDonald House Charities of Chicagoland & Northwest Indiana

    - 1 year

    Social Services

    Helped prepare meals for families staying at the Chicago Ronald McDonald House.

Projects

  • Project HERO (Highlighting Effective Rapid Opportunities)

    -

    Our sales pipeline is ripe with opportunity for advanced analytics and data science. Our mission has two parts: a) define 'rapid' win, b) predict current in-pipeline opportunities that are likely to be a rapid win.

    Identifying rapid win is a necessary building block to generating predictions because we needed a principled definition around 'rapid', and we needed labels in which to generate predictions on. 'Rapid' was determined by analysis of the time-to-win distribution of our…

    Our sales pipeline is ripe with opportunity for advanced analytics and data science. Our mission has two parts: a) define 'rapid' win, b) predict current in-pipeline opportunities that are likely to be a rapid win.

    Identifying rapid win is a necessary building block to generating predictions because we needed a principled definition around 'rapid', and we needed labels in which to generate predictions on. 'Rapid' was determined by analysis of the time-to-win distribution of our historical won-sales opportunities.

    With labels in hand, a supervised model was developed to predict 'rapid' or 'not-rapid' labels on current pipeline opportunities. Further enabling smarter, better business actions, we highlighted the independent variables that had the greatest impact on predicting 'rapid' labels -- providing the business user with a clear understanding of which levers to pull.

    Technologies used:
    * SQL
    * R (tidyverse, fitdistrplus)
    * python (pyodbc, pandas, scikit learn, statsmodels)
    * jupyter notebook

    Techniques used:
    * data wrangling
    * probability distribution fitting
    * supervised machine learning

    Other creators
  • Project OREO (Overlapping Repop Engagement Opportunities)

    -

    There was a need to identify accounts with aging scientific instruments that may require replacement. While many frameworks and analytical products were in place, we identified an opportunity to enhance the current-state with a composite likelihood-to-buy score (based on historical purchase behavior). We extended our recommendation to prescribe under-purchased consumables for those instruments using unsupervised clustering methods, providing within-groups context for expected consumable…

    There was a need to identify accounts with aging scientific instruments that may require replacement. While many frameworks and analytical products were in place, we identified an opportunity to enhance the current-state with a composite likelihood-to-buy score (based on historical purchase behavior). We extended our recommendation to prescribe under-purchased consumables for those instruments using unsupervised clustering methods, providing within-groups context for expected consumable purchases.

    This product increased our monthly conversion rate by 200%.

    Technologies used:
    * SQL
    * Python (pyodbc, pandas, scikit learn, statsmodels)
    * PowerBI

    Techniques used:
    * data wrangling
    * modified RFM (recency, frequency, monetary)
    * unsupervised machine learning
    * data visualization

    Other creators
  • Project SASSY (Selective Account Sizing & Segmentation Yield)

    -

    Using transactional and customer attribute data assets to segment customers for recommending right-sized support packages for their recently purchased equipment. Techniques include unsupervised machine learning methods and robust engineering principles.

    Other creators
  • Project SRI (Supplier Risk Identifier)

    -

    A $5bn, national food distributor relied heavily on manual, excel-based reporting processes to monitor high debt-risk suppliers. This limited the efficacy of KPI measurement due to poor measurement methods, infrequent analysis and inability to forecast accurately. A solution was designed and implemented to automate data preparation and reporting, reliably identify suppliers at-risk of insolvency, and forecast next quarter debit balance.

    Technologies used:
    * SQL
    * Python (pyodbc,…

    A $5bn, national food distributor relied heavily on manual, excel-based reporting processes to monitor high debt-risk suppliers. This limited the efficacy of KPI measurement due to poor measurement methods, infrequent analysis and inability to forecast accurately. A solution was designed and implemented to automate data preparation and reporting, reliably identify suppliers at-risk of insolvency, and forecast next quarter debit balance.

    Technologies used:
    * SQL
    * Python (pyodbc, pandas, scikit learn, statsmodels)
    * PowerBI

    Techniques used:
    * data wrangling
    * time series forecasting
    * supervised machine learning
    * data visualization

    Other creators

Recommendations received

8 people have recommended Vincent

Join now to view

More activity by Vincent

View Vincent’s full profile

  • See who you know in common
  • Get introduced
  • Contact Vincent directly
Join to view full profile

Other similar profiles

Explore collaborative articles

We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.

Explore More

Others named Vincent Chung in United States