This data science project investigates pricing behaviour in the pharmaceutical industry by analysing prescription drug pricing data between 2019 and 2020. The analysis focuses on Wholesale Acquisition Cost (WAC), a pricing benchmark used within the US pharmaceutical market to represent the list price set by manufacturers before rebates or discounts.

Using Power BI and structured data science methods, the project explores how drug prices change after market introduction and identifies manufacturers or products that demonstrate unusually high price inflation. By transforming raw pharmaceutical datasets into structured analytical models, the project highlights patterns in pricing behaviour and potential risks within drug pricing strategies.

Data Science Workflow

The project followed a complete data science pipeline including data preparation, transformation, modelling, and visual analytics.

Multiple pharmaceutical pricing datasets were merged and cleaned using Power Query, creating a consolidated dataset called Drug_Prices_Master_WAC. Data preparation involved removing duplicates, filtering invalid records, standardising column formats, and converting pricing data into numerical formats suitable for analysis.

The final dataset was structured using a star schema data model, linking a central pricing fact table with supporting dimension tables such as date information and historical drug reports. This structure enabled efficient querying and interactive dashboard exploration.

Feature Engineering and Metrics

To evaluate pricing behaviour, several analytical variables were created using DAX calculations.

These included:

• WAC Difference – measures the absolute price change from launch price to current price
• WAC Ratio – compares the current drug price to its original introduction price
• WAC % Increase – measures percentage price growth over time
• Annualised Growth Rate – calculates yearly price inflation for each drug
• Volatility Score – measures manufacturer-level pricing variability using statistical deviation

Additional classification metrics such as High Risk Drug and WAC Growth Category were created to categorise drugs into pricing risk groups including stable, moderate growth, high growth, and extreme increase. Cost Transparency in Pharmaceut…

These engineered features allowed the analysis to move beyond simple price comparisons and instead detect pricing behaviour patterns across manufacturers and time periods.

Dashboard Analytics

Interactive dashboards were built in Power BI to communicate insights from the dataset.

The main dashboard presents a high-level overview of the pharmaceutical market within the dataset, highlighting:

• 440 drugs analysed
• 93 manufacturers
• 21 high-risk drugs with extreme price increases

Time-series visualisations track drug introductions over time, while additional dashboards explore pricing behaviour across manufacturers and individual drugs.

Manufacturer-level analysis revealed that the mean WAC introduction price was $17,140 while the median price was only $2,040, indicating strong skewness caused by extreme outliers in the pharmaceutical market.

Key Findings

Several important insights emerged from the analysis.

The dataset revealed 21 drugs classified as high risk due to extreme price inflation, with some drugs increasing to several times their original introduction price.

One notable outlier identified in the dataset was Lexicon Pharmaceuticals, with a reported drug launch price exceeding $2.1 million, highlighting the extreme variability present in pharmaceutical pricing strategies.

The analysis also showed that pricing behaviour varies significantly across manufacturers, with some companies demonstrating relatively stable pricing while others exhibited volatile or aggressive price increases.

Policy and Ethical Context

Beyond technical analysis, the project situates these findings within broader healthcare policy discussions surrounding drug pricing transparency.

Wholesale Acquisition Cost increases have become a key concern in US healthcare policy, with legislation such as the Inflation Reduction Act (2022) enabling Medicare to negotiate prices for certain high-cost drugs.

By identifying drugs with extreme pricing growth, the analytical framework developed in this project could support regulatory oversight and policy evaluation related to drug affordability.

Conclusion

This project demonstrates how data science and visual analytics can be used to analyse complex pharmaceutical datasets and uncover patterns in drug pricing behaviour.

Through data transformation, feature engineering, and interactive dashboards, the analysis provides a framework for identifying high-risk pricing patterns within the pharmaceutical industry. The findings highlight how analytical tools can support evidence-based discussions around healthcare transparency, affordability, and regulatory oversight.