Top 24 Data Science Case Studies [2026]

Team DigitalDefynd

In an era where data is the new gold, harnessing its power through data science has led to groundbreaking advancements across industries. From personalized marketing to predictive maintenance, the applications of data science are not only diverse but transformative. This compilation of the top 25 data science case studies showcases the profound impact of intelligent data utilization in solving real-world problems. These examples span various sectors, including healthcare, finance, transportation, and manufacturing, illustrating how data-driven decisions shape business operations’ future, enhance efficiency, and optimize user experiences. As we delve into these case studies, we witness the incredible potential of data science to innovate and drive success in today’s data-centric world.

Top 24 Data Science Case Studies [2026]

Case Study 1 – Personalized Marketing (Amazon)

Challenge: Amazon aimed to enhance user engagement by tailoring product recommendations to individual preferences, requiring the real-time processing of vast data volumes.

Solution: Amazon implemented a sophisticated machine learning algorithm known as collaborative filtering, which analyzes users’ purchase history, cart contents, product ratings, and browsing history, along with the behavior of similar users. This approach enables Amazon to offer highly personalized product suggestions.

Overall Impact:

Increased Customer Satisfaction: Tailored recommendations improved the shopping experience.
Higher Sales Conversions: Relevant product suggestions boosted sales.

Key Takeaways:

Personalized Marketing Significantly Enhances User Engagement: Demonstrating how tailored interactions can deepen user involvement and satisfaction.
Effective Use of Big Data and Machine Learning Can Transform Customer Experiences: These technologies redefine the consumer landscape by continuously adapting recommendations to changing user preferences and behaviors.

This strategy has proven pivotal in increasing Amazon’s customer loyalty and sales by making the shopping experience more relevant and engaging.

Case Study 2 – Real-Time Pricing Strategy (Uber)

Challenge: Uber needed to adjust its pricing dynamically to reflect real-time demand and supply variations across different locations and times, aiming to optimize driver incentives and customer satisfaction without manual intervention.

Solution: Uber introduced a dynamic pricing model called “surge pricing.” This system uses data science to automatically calculate fares in real time based on current demand and supply data. The model incorporates traffic conditions, weather forecasts, and local events to adjust prices appropriately.

Overall Impact:

Optimized Ride Availability: The model reduced customer wait times by incentivizing more drivers to be available during high-demand periods.
Increased Driver Earnings: Drivers benefitted from higher earnings during surge periods, aligning their incentives with customer demand.

Key Takeaways:

Efficient Balance of Supply and Demand: Dynamic pricing matches ride availability with customer needs.
Importance of Real-Time Data Processing: The real-time processing of data is crucial for responsive and adaptive service delivery.

Uber’s implementation of surge pricing illustrates the power of using real-time data analytics to create a flexible and responsive pricing system that benefits both consumers and service providers, enhancing overall service efficiency and satisfaction.

Case Study 3 – Fraud Detection in Banking (JPMorgan Chase)

Challenge: JPMorgan Chase faced the critical need to enhance its fraud detection capabilities to safeguard the institution and its customers from financial losses. The primary challenge was detecting fraudulent transactions swiftly and accurately in a vast stream of legitimate banking activities.

Solution: The bank implemented advanced machine learning models that analyze real-time transaction patterns and customer behaviors. These models are continuously trained on vast amounts of historical fraud data, enabling them to identify and flag transactions that significantly deviate from established patterns, which may indicate potential fraud.

Overall Impact:

Substantial Reduction in Fraudulent Transactions: The advanced detection capabilities led to a marked decrease in fraud occurrences.
Enhanced Security for Customer Accounts: Customers experienced greater security and trust in their transactions.

Key Takeaways:

Effectiveness of Machine Learning in Fraud Detection: Machine learning models are greatly effective at identifying fraud activities within large datasets.
Importance of Ongoing Training and Updates: Continuous training and updating of models are crucial to adapt to evolving fraudulent techniques and maintain detection efficacy.

JPMorgan Chase’s use of machine learning for fraud detection demonstrates how financial institutions can leverage advanced analytics to enhance security measures, protect financial assets, and build customer trust in their banking services.

Case Study 4 – Optimizing Healthcare Outcomes (Mayo Clinic)

Challenge: The Mayo Clinic aimed to enhance patient outcomes by predicting diseases before they reach critical stages. This involved analyzing large volumes of diverse data, including historical patient records and real-time health metrics from various sources like lab results and patient monitors.

Solution: The Mayo Clinic employed predictive analytics to integrate and analyze this data to build models that predict patient risk for diseases such as diabetes and heart disease, enabling earlier and more targeted interventions.

Overall Impact:

Improved Patient Outcomes: Early identification of at-risk patients allowed for timely medical intervention.
Reduction in Healthcare Costs: Preventing disease progression reduces the need for more extensive and costly treatments later.

Key Takeaways:

Early Identification of Health Risks: Predictive models are essential for identifying at-risk patients early, improving the chances of successful interventions.
Integration of Multiple Data Sources: Combining historical and real-time data provides a comprehensive view that enhances the accuracy of predictions.

Case Study 5 – Streamlining Operations in Manufacturing (General Electric)

Challenge: General Electric needed to optimize its manufacturing processes to reduce costs and downtime by predicting when machines would likely require maintenance to prevent breakdowns.

Solution: GE leveraged data from sensors embedded in machinery to monitor their condition continuously. Data science algorithms analyze this sensor data to predict when a machine is likely to disappoint, facilitating preemptive maintenance and scheduling.

Overall Impact:

Reduction in Unplanned Machine Downtime: Predictive maintenance helped avoid unexpected breakdowns.
Lower Maintenance Costs and Improved Machine Lifespan: Regular maintenance based on predictive data reduced overall costs and extended the life of machinery.

Key Takeaways:

Predictive Maintenance Enhances Operational Efficiency: Using data-driven predictions for maintenance can significantly reduce downtime and operational costs.
Value of Sensor Data: Continuous monitoring and data analysis are crucial for forecasting equipment health and preventing failures.

Case Study 6 – Enhancing Supply Chain Management (DHL)

Challenge: DHL sought to optimize its global logistics and supply chain operations to decreases expenses and enhance delivery efficiency. It required handling complex data from various sources for better route planning and inventory management.

Solution: DHL implemented advanced analytics to process and analyze data from its extensive logistics network. This included real-time tracking of shipments, analysis of weather conditions, traffic patterns, and inventory levels to optimize route planning and warehouse operations.

Overall Impact:

Enhanced Efficiency in Logistics Operations: More precise route planning and inventory management improved delivery times and reduced resource wastage.
Reduced Operational Costs: Streamlined operations led to significant cost savings across the supply chain.

Key Takeaways:

Critical Role of Comprehensive Data Analysis: Effective supply chain management depends on integrating and analyzing data from multiple sources.
Benefits of Real-Time Data Integration: Real-time data enhances logistical decision-making, leading to more efficient and cost-effective operations.

Case Study 7 – Predictive Maintenance in Aerospace (Airbus)

Challenge: Airbus faced the challenge of predicting potential failures in aircraft components to enhance safety and reduce maintenance costs. The key was to accurately forecast the lifespan of parts under varying conditions and usage patterns, which is critical in the aerospace industry where safety is paramount.

Solution: Airbus tackled this challenge by developing predictive models that utilize data collected from sensors installed on aircraft. These sensors continuously monitor the condition of various components, providing real-time data that the models analyze. The predictive algorithms assess the likelihood of component failure, enabling maintenance teams to schedule repairs or replacements proactively before actual failures occur.

Overall Impact:

Increased Safety: The ability to predict and prevent potential in-flight failures has significantly improved the safety of Airbus aircraft.
Reduced Costs: By optimizing maintenance schedules and minimizing unnecessary checks, Airbus has been able to cut down on maintenance expenses and reduce aircraft downtime.

Key Takeaways:

Enhanced Safety through Predictive Analytics: The use of predictive analytics in monitoring aircraft components plays a crucial role in preventing failures, thereby enhancing the overall safety of aviation operations.
Valuable Insights from Sensor Data: Real-time data from operational use is critical for developing effective predictive maintenance strategies. This data provides insights for understanding component behavior under various conditions, allowing for more accurate predictions.

This case study demonstrates how Airbus leverages advanced data science techniques in predictive maintenance to ensure higher safety standards and more efficient operations, setting an industry benchmark in the aerospace sector.

Case Study 8 – Enhancing Film Recommendations (Netflix)

Challenge: Netflix aimed to improve customer retention and engagement by enhancing the accuracy of its recommendation system. This task involved processing and analyzing vast amounts of data to understand diverse user preferences and viewing habits.

Solution: Netflix employed collaborative filtering techniques, analyzing user behaviors (like watching, liking, or disliking content) and similarities between content items. This data-driven approach allows Netflix to refine and personalize recommendations continuously based on real-time user interactions.

Overall Impact:

Increased Viewer Engagement: Personalized recommendations led to longer viewing sessions.
Higher Customer Satisfaction and Retention Rates: Tailored viewing experiences improved overall customer satisfaction, enhancing loyalty.

Key Takeaways:

Tailoring User Experiences: Machine learning is pivotal in personalizing media content, significantly impacting viewer engagement and satisfaction.
Importance of Continuous Updates: Regularly updating recommendation algorithms is essential to maintain relevance and effectiveness in user engagement.

Case Study 9 – Traffic Flow Optimization (Google)

Challenge: Google needed to optimize traffic flow within its Google Maps service to reduce congestion and improve routing decisions. This required real-time analysis of extensive traffic data to predict and manage traffic conditions accurately.

Solution: Google Maps integrates data from multiple sources, including satellite imagery, sensor data, and real-time user location data. These data points are used to model traffic patterns and predict future conditions dynamically, which informs updated routing advice.

Overall Impact:

Reduced Traffic Congestion: More efficient routing reduced overall traffic buildup.
Enhanced Accuracy of Traffic Predictions and Routing: Improved predictions led to better user navigation experiences.

Key Takeaways:

Integration of Multiple Data Sources: Combining various data streams enhances the accuracy of traffic management systems.
Advanced Modeling Techniques: Sophisticated models are crucial for accurately predicting traffic patterns and optimizing routes.

Case Study 10 – Risk Assessment in Insurance (Allstate)

Challenge: Allstate sought to refine its risk assessment processes to offer more accurately priced insurance products, challenging the limitations of traditional actuarial models through more nuanced data interpretations.

Solution: Allstate enhanced its risk assessment framework by integrating machine learning, allowing for granular risk factor analysis. This approach utilizes individual customer data such as driving records, home location specifics, and historical claim data to tailor insurance offerings more accurately.

Overall Impact:

More Precise Risk Assessment: Improved risk evaluation led to more tailored insurance offerings.
Increased Market Competitiveness: Enhanced pricing accuracy boosted Allstate’s competitive edge in the insurance market.

Key Takeaways:

Nuanced Understanding of Risk: Machine learning provides a deeper, more nuanced understanding of risk than traditional models, leading to better risk pricing.
Personalized Pricing Strategies: Leveraging detailed customer data in pricing strategies enhances customer satisfaction and business performance.

Case Study 11 – Energy Consumption Reduction (Google DeepMind)

Challenge: Google DeepMind aimed to significantly reduce the high energy consumption required for cooling Google’s data centers, which are crucial for maintaining server performance but also represent a major operational cost.

Solution: DeepMind implemented advanced AI algorithms to optimize the data center cooling systems. These algorithms predict temperature fluctuations and adjust cooling processes accordingly, saving energy and reducing equipment wear and tear.

Overall Impact:

Reduction in Energy Consumption: Achieved a 40% reduction in energy used for cooling.
Decrease in Operational Costs and Environmental Impact: Lower energy usage resulted in cost savings and reduced environmental footprint.

Key Takeaways:

AI-Driven Optimization: AI can significantly decrease energy usage in large-scale infrastructure.
Operational Efficiency Gains: Efficiency improvements in operational processes lead to cost savings and environmental benefits.

Case Study 12 – Improving Public Safety (New York City Police Department)

Challenge: The NYPD needed to enhance its crime prevention strategies by better predicting where and when crimes were most likely to occur, requiring sophisticated analysis of historical crime data and environmental factors.

Solution: The NYPD implemented a predictive policing system that utilizes data analytics to identify potential crime hotspots based on trends and patterns in past crime data. Officers are preemptively dispatched to these areas to deter criminal activities.

Overall Impact:

Reduction in Crime Rates: There is a notable decrease in crime in areas targeted by predictive policing.
More Efficient Use of Police Resources: Enhanced allocation of resources where needed.

Key Takeaways:

Effectiveness of Data-Driven Crime Prevention: Targeting resources based on data analytics can significantly reduce crime.
Proactive Law Enforcement: Predictive analytics enable a shift from reactive to proactive law enforcement strategies.

Case Study 13 – Enhancing Agricultural Yields (John Deere)

Challenge: John Deere aimed to help farmers increase agricultural productivity and sustainability by optimizing various farming operations from planting to harvesting.

Solution: Utilizing data from sensors on equipment and satellite imagery, John Deere developed algorithms that provide actionable insights for farmers on optimal planting times, water usage, and harvest schedules.

Overall Impact:

Increased Crop Yields: More efficient farming methods led to higher yields.
Enhanced Sustainability of Farming Practices: Improved resource management contributed to more sustainable agriculture.

Key Takeaways:

Precision Agriculture: Significantly improves productivity and resource efficiency.
Data-Driven Decision-Making: Enables better farming decisions through timely and accurate data.

Case Study 14 – Streamlining Drug Discovery (Pfizer)

Challenge: Pfizer faced the need to accelerate the process of discoverying drug and improve the success rates of clinical trials.

Solution: Pfizer employed data science to simulate and predict outcomes of drug trials using historical data and predictive models, optimizing trial parameters and improving the selection of drug candidates.

Overall Impact:

Accelerated Drug Development: Reduced time to market for new drugs.
Increased Efficiency and Efficacy in Clinical Trials: More targeted trials led to better outcomes.

Key Takeaways:

Reduction in Drug Development Time and Costs: Data science streamlines the R&D process.
Improved Clinical Trial Success Rates: Predictive modeling enhances the accuracy of trial outcomes.

Case Study 15 – Media Buying Optimization (Procter & Gamble)

Challenge: Procter & Gamble aimed to maximize the ROI of their extensive advertising budget by optimizing their media buying strategy across various channels.

Solution: P&G analyzed extensive data on consumer behavior and media consumption to identify the most effective times and channels for advertising, allowing for highly targeted ads that reach the intended audience at optimal times.

Overall Impact:

Improved Effectiveness of Advertising Campaigns: More effective ads increased campaign impact.
Increased Sales and Better Budget Allocation: Enhanced ROI from more strategic media spending.

Key Takeaways:

Enhanced Media Buying Strategies: Data analytics significantly improves media buying effectiveness.
Insights into Consumer Behavior: Understanding consumer behavior is crucial for optimizing advertising ROI.

Case Study 16 – Enhancing E-commerce Customer Experience with AI (Zalando)

Challenge: Zalando aimed to enhance the online shopping experience by improving the accuracy of size recommendations, a common issue that leads to high return rates in online apparel shopping.

Solution: Zalando developed an AI-driven size recommendation engine that analyzes past purchase and return data in combination with customer feedback and preferences. This system utilizes machine learning to predict the best-fit size for customers based on their unique body measurements and purchase history.

Overall Impact:

Reduced Return Rates: More accurate size recommendations decreased the returns due to poor fit.
Improved Customer Satisfaction: Customers experienced a more personalized shopping journey, enhancing overall satisfaction.

Key Takeaways:

Customization Through AI: Personalizing customer experience can significantly impact satisfaction and business metrics.
Data-Driven Decision-Making: Utilizing customer data effectively can improve business outcomes by reducing costs and enhancing the user experience.

Case Study 17 – Optimizing Energy Grid Performance with Machine Learning (Enel Group)

Challenge: Enel Group, one of the largest power companies, faced challenges in managing and optimizing the performance of its vast energy grids. The primary goal was to increase the efficiency of energy distribution and reduce operational costs while maintaining reliability in the face of fluctuating supply and demand.

Solution: Enel Group implemented a machine learning-based system that analyzes real-time data from smart meters, weather stations, and IoT devices across the grid. This system is designed to predict peak demand times, potential outages, and equipment failures before they occur. By integrating these predictions with automated grid management tools, Enel can dynamically adjust energy flows, allocate resources more efficiently, and schedule maintenance proactively.

Overall Impact:

Enhanced Grid Efficiency: Improved distribution management, reduced energy wastage, and optimized resource allocation.
Reduced Operational Costs: Predictive maintenance and better grid management decreased the frequency and cost of repairs and outages.

Key Takeaways:

Predictive Maintenance in Utility Networks: Advanced analytics can preemptively identify issues, saving costs and enhancing service reliability.
Real-Time Data Integration: Leveraging data from various sources in real-time enables more agile and informed decision-making in energy management.

Case Study 18 – Personalizing Movie Streaming Experience (WarnerMedia)

Challenge: WarnerMedia sought to enhance viewer engagement and subscription retention rates on its streaming platforms by providing more personalized content recommendations.

Solution: WarnerMedia deployed a sophisticated data science strategy, utilizing deep learning algorithms to analyze viewer behaviors, including viewing history, ratings given to shows and movies, search patterns, and demographic data. This analysis helped create highly personalized viewer profiles, which were then used to tailor content recommendations, homepage layouts, and promotional offers specifically to individual preferences.

Overall Impact:

Increased Viewer Engagement: Personalized recommendations resulted in extended viewing times and increased interactions with the platform.
Higher Subscription Retention: Tailored user experiences improved overall satisfaction, leading to lower churn rates.

Key Takeaways:

Deep Learning Enhances Personalization: Deep learning algorithms allow a more nuanced knowledge of consumer preferences and behavior.
Data-Driven Customization is Key to User Retention: Providing a customized experience based on data analytics is critical for maintaining and growing a subscriber base in the competitive streaming market.

Case Study 19- Improving Online Retail Sales through Customer Sentiment Analysis (Zappos)

Challenge: Zappos, an online shoe and clothing retailer, aimed to enhance customer satisfaction and boost sales by better understanding customer sentiments and preferences across various platforms.

Solution: Zappos implemented a comprehensive sentiment analysis program that utilized natural language processing (NLP) techniques to gather and analyze customer feedback from social media, product reviews, and customer support interactions. This data was used to identify emerging trends, customer pain points, and overall sentiment towards products and services. The insights derived from this analysis were subsequently used to customize marketing strategies, enhance product offerings, and improve customer service practices.

Overall Impact:

Enhanced Product Selection and Marketing: Insight-driven adjustments to inventory and marketing strategies increased relevancy and customer satisfaction.
Improved Customer Experience: By addressing customer concerns and preferences identified through sentiment analysis, Zappos enhanced its overall customer service, increasing loyalty and repeat business.

Key Takeaways:

Power of Sentiment Analysis in Retail: Understanding and reacting to customer emotions and opinions can significantly impact sales and customer satisfaction.
Strategic Use of Customer Feedback: Leveraging customer feedback to drive business decisions helps align product offerings and services with customer expectations, fostering a positive brand image.

Case Study 20 – Streamlining Airline Operations with Predictive Analytics (Delta Airlines)

Challenge: Delta Airlines faced operational challenges, including flight delays, maintenance scheduling inefficiencies, and customer service issues, which impacted passenger satisfaction and operational costs.

Solution: Delta implemented a predictive analytics system that integrates data from flight operations, weather reports, aircraft sensor data, and historical maintenance records. The system predicts potential delays using machine learning models and suggests optimal maintenance scheduling. Additionally, it forecasts passenger load to optimize staffing and resource allocation at airports.

Overall Impact:

Reduced Flight Delays: Predictive insights allowed for better planning and reduced unexpected delays.
Enhanced Maintenance Efficiency: Maintenance could be scheduled proactively, decreasing the time planes spend out of service.
Improved Passenger Experience: With better resource management, passenger handling became more efficient, enhancing overall customer satisfaction.

Key Takeaways:

Operational Efficiency Through Predictive Analytics: Leveraging data for predictive purposes significantly improves operational decision-making.
Data Integration Across Departments: Coordinating data from different sources provides a holistic view crucial for effective airline management.

Case Study 21 – Enhancing Financial Advisory Services with AI (Morgan Stanley)

Challenge: Morgan Stanley sought to offer clients more personalized and effective financial guidance. The challenge was seamlessly integrating vast financial data with individual client profiles to deliver tailored investment recommendations.

Solution: Morgan Stanley developed an AI-powered platform that utilizes natural language processing and ML to analyze financial markets, client portfolios, and historical investment performance. The system identifies patterns and predicts market trends while considering each client’s financial goals, risk tolerance, and investment history. This integrated approach enables financial advisors to offer highly customized advice and proactive investment strategies.

Overall Impact:

Improved Client Satisfaction: Clients received more relevant and timely investment recommendations, enhancing their overall satisfaction and trust in the advisory services.
Increased Efficiency: Advisors were able to manage client portfolios more effectively, using AI-driven insights to make faster and more informed decisions.

Key Takeaways:

Personalization through AI: Advanced analytics and AI can significantly enhance the personalization of financial services, leading to better client engagement.
Data-Driven Decision Making: Leveraging diverse data sets provides a comprehensive understanding crucial for tailored financial advising.

Case Study 22 – Optimizing Inventory Management in Retail (Walmart)

Challenge: Walmart sought to improve inventory management across its vast network of stores and warehouses to reduce overstock and stockouts, which affect customer satisfaction and operational efficiency.

Solution: Walmart implemented a robust data analytics system that integrates real-time sales data, supply chain information, and predictive analytics. This system uses machine learning algorithms to forecast demand for thousands of products at a granular level, considering factors such as seasonality, local events, and economic trends. The predictive insights allow Walmart to dynamically adjust inventory levels, optimize restocking schedules, and manage distribution logistics more effectively.

Overall Impact:

Reduced Inventory Costs: More accurate demand forecasts helped minimize overstock and reduce waste.
Enhanced Customer Satisfaction: Improved stock availability led to better in-store experiences and higher customer satisfaction.

Key Takeaways:

Precision in Demand Forecasting: Advanced data analytics and machine learning significantly enhance demand forecasting accuracy in retail.
Integrated Data Systems: Combining various data sources provides a comprehensive view of inventory needs, improving overall supply chain efficiency.

Case Study 23: Enhancing Network Security with Predictive Analytics (Cisco)

Challenge: Cisco encountered difficulties protecting its extensive network infrastructure from increasingly complex cyber threats. The objective was to bolster their security protocols by anticipating potential breaches before they happen.

Solution: Cisco developed a predictive analytics solution that leverages ML algorithms to analyze patterns in network traffic and identify anomalies that could suggest a security threat. By integrating this system with their existing security protocols, Cisco can dynamically adjust defenses and alert system administrators about potential vulnerabilities in real-time.

Overall Impact:

Improved Security Posture: The predictive system enabled proactive responses to potential threats, significantly reducing the incidence of successful cyber attacks.
Enhanced Operational Efficiency: Automating threat detection and response processes allowed Cisco to manage network security more efficiently, with fewer resources dedicated to manual monitoring.

Key Takeaways:

Proactive Security Measures: Employing predictive cybersecurity analytics helps organizations avoid potential threats.
Integration of Machine Learning: Machine learning is crucial for effectively detecting patterns and anomalies that human analysts might overlook, leading to stronger security measures.

Case Study 24 – Improving Agricultural Efficiency with IoT and AI (Bayer Crop Science)

Challenge: Bayer Crop Science aimed to enhance agricultural efficiency and crop yields for farmers worldwide, facing the challenge of varying climatic conditions and soil types that affect crop growth differently.

Solution: Bayer deployed an integrated platform that merges IoT sensors, satellite imagery, and AI-driven analytics. This platform gathers real-time weather conditions, soil quality, and crop health data. Utilizing machine learning models, the system processes this data to deliver precise agricultural recommendations to farmers, including optimal planting times, watering schedules, and pest management strategies.

Overall Impact:

Increased Crop Yields: Tailored agricultural practices led to higher productivity per hectare.
Reduced Resource Waste: Efficient water use, fertilizers, and pesticides minimized environmental impact and operational costs.

Key Takeaways:

Precision Agriculture: Leveraging IoT and AI enables more precise and data-driven agricultural practices, enhancing yield and efficiency.
Sustainability in Farming: Advanced data analytics enhance the sustainability of farming by optimizing resource utilization and minimizing waste.

Conclusion

The power of data science in transforming industries is undeniable, as demonstrated by these 25 compelling case studies. Through the strategic application of machine learning, predictive analytics, and AI, companies are solving complex challenges and gaining a competitive edge. The insights gleaned from these cases highlight the critical role of data science in enhancing decision-making processes, improving operational efficiency, and elevating customer satisfaction. As we look to the future, the role of data science is set to grow, promising even more innovative solutions and smarter strategies across all sectors. These case studies inspire and serve as a roadmap for harnessing the transformative power of data science in the journey toward digital transformation.