Team members:
Jalpesh Chavda, Associate Director, with:
- Trinath Chittiprolu
- Shashank Vishwakarma
- Shachi Sankhla
- Apeksha Lanjewar
- Suman Dewan
Country: United States
Organization: Novartis Pharmaceuticals
Novartis’s purpose is to reimagine medicine to improve and extend people's lives. We use innovative science and technology to address some of society's most challenging healthcare issues. We discover and develop breakthrough treatments and find new ways to deliver them to as many people as possible. We also aim to reward those who invest their money, time, and ideas in our company.
Awards Categories:
- Best Acceleration Use Case
- Best Approach for Building Trust in AI
Business Challenge:
The business team has a weekly task of updating data in Excel to generate important metrics. This process involves repetitive manual calculations of various key performance metrics and decisions are made based on the outcomes.
However, the team is currently facing some obstacles such as modifying the breadth (active writers) and depth (prescription per writer) parameters in the existing process. Additionally, there is a lack of real-time data refresh and ineffective data tracking, leading to discrepancies due to human error. This is also affecting the team's ability to identify risks in budget forecasts and field-force allocation.
To solve these issues, our Data Engineer and Data Science team came together and developed an automated solution using Dataiku. We used SARIMAX in Python for time series modeling to forecast future values based on historical data. With this solution, the team can now avoid repetitive manual calculations and make more informed decisions based on accurate and real-time data.
Business Solution:
Our company uses Dataiku, which offers various modular components for project design, batch process automation, real-time API scoring, and AI governance capabilities.
With the help of Dataiku, we created a customized simulation dashboard for our client, which automates the analysis of budget allocation and field-force allotments, allowing them to make better decisions for demand-growth opportunities.
Dataiku has allowed us to create customized machine learning forecast models and present the results in a user-friendly environment, highlighting the predictions for the most practical scenarios to reach business goals. It also has informative wikis for all the data sources and processes, and supports usage of global variables and parameterization not only for the table details but also for run-time flexibility.
Day-to-day Change:
Adopting Dataiku was the natural next step for our team. It exposed us to new functionalities which are now also being adopted by other teams.
Dataiku provides self-explanatory workflows, dividing each workflow into multiple zones to reflect the pipeline. This allows different team members to work on various parts of the project simultaneously, thus reducing implementation time.
It also has informative wikis for all the data sources and processes, and supports usage of global variables and parameterization not only for the table details but also for run-time flexibility. This cuts down on the endless rabbit hole of users seeking support, now we have a one-stop shop for all our questions.
Finally, Dataiku cuts down on the total execution time thanks to its parallel execution of scenarios and workflow schedulers, reducing the overall runtime of various tedious processes. This allows us to have better deliverables in less time, making us more productive.
Business Area Enhanced: Marketing/Sales/Customer Relationship Management
Use Case Stage: In Production
Value Generated:
Dataiku has a pushdown design that ensures optimum performance and efficiency, even when handling huge workloads. This allows enterprises to use current, elastic, and highly scalable computing platforms like SQL databases, Spark, Kubernetes, and others.
Additionally, the metrics, checks, and testing capabilities provided by Dataiku have enabled us to add quality assurance to our models.
Dataiku offers a reliable auto-documentation feature that enables organizations to maintain consistent records of their AI projects. This helps them comply with regulations, and teams can save time by not having to spend countless hours creating project documents or worrying about additional IT maintenance. With minimal manual intervention, this feature provides increased efficiency and value.
Value Brought by Dataiku:
Our team opted to use Dataiku as our solution implementation tool because of its numerous features that are truly groundbreaking for the industry. With Dataiku, we can automate our machine learning model solution process using built-in ETL capabilities. This enables us to create customized machine learning forecast models that are presented in a user-friendly environment, highlighting the predictions for practical scenarios that help us reach our business goals.
One of the things that attract us most to Dataiku is the automation features built into the application, such as real-time data refresh and analysis, automatic periodic reports refresh, and pre-built visualizations and reporting templates. Most importantly, we appreciate the ability to track actual versus forecast variance using simple and easy-to-understand dashboards and sliders.
These are the key features of the solution we selected from Dataiku:
- Automated data feeds for real-time data analysis
- KPI reports that refresh automatically with monthly data updates
- Pre-built templates for visualization and reporting
- Customized Forecast Model based on business needs, including Trends, Seasonality, and Error
- Ability to track variances between actual and forecasted
Value Type:
- Reduce cost
- Save time
- Increase trust
Disclaimer: The views and opinions expressed in this document are those of the submitters and do not necessarily reflect the official policy or position of Novartis or any of its officers.
