Braskem - Historical Data Access Democratization
Team members:
- Luis Paulo Bernardi
- Jorge Manuel Jardim da Silva
- João Vitor da Silva Bastos
Country: Brazil
Organization: Braskem
Braskem is a global company with industrial units located in Brazil, the United States, Mexico, Europa and Asia. Founded in 2002 via the integration of six companies from the Odebrecht Organization and the Mariani Group, we are the sixth largest petrochemical company in the world in the production of thermoplastic resins. With customers in more than 71 countries, we are the market leader in the Americas and pioneers in the production of biopolymers (plastic made of renewable raw material) on an industrial scale.
Awards Category:
- Best Data Democratization Program
Business Challenge:
The I&T area has several laboratories that generate a large amount of data on a daily basis. As a result, there is a large amount of historical data of various types, such as different file formats, texts, images, and others. Because of this diversity, navigating the databases may not be very intuitive for most users, as knowledge of SQL and data structure is not common across all teams. In addition, using spreadsheets for large amounts of data isn't very practical.
Non-intuitive access to data can result in data not being used to its full potential or requests for analyses that have already been performed. This results in additional costs, in addition to waiting for data that could be available immediately.
The democratization of data access should allow all I&T members to access historical data in an easy and intuitive way. With the ability to access databases and create and publish web applications within the company in a relatively simple and secure way, Dataiku was identified as a great tool to meet this need.
Business Solution:
Since there is a wide variety of data in the database, a project was created in Dataiku to aggregate the web applications. Specific webapps were developed for each type of data, as visualization and comparison are associated with different types of data. The transformation and processing of the datasets that feed the web applications were performed by different zones within the flow.
In short, each web application was developed as follows:
- Conceptually design the visualization for each type of data;
- Connecting to the database, loading, processing and generating the dataset;
- Build the web application;
- Gathering user insights and improving the data visualization;
- Publish the web application;
As a result, access to data was greatly democratized and visualization of historical data was greatly facilitated. In addition, users benefited from being able to compare data directly within the search and view itself, which was not previously possible.
Another benefit that Dataiku brought was the centralization of this development, as it allows the centralization of all data transformation processes and web application code. This simplified management and access.
Day-to-day Change:
Previously, democratizing access to data required the development of spreadsheets, file sharing, or the use of BI tools that did not have all the functionality needed to process and visualize different types of scientific data. Now, Dataiku provides a common platform for accessing data, manipulating it in a simple way, and creating web applications developed in Python.
What used to take a lot of effort can now be done on a common platform and with a tooling that allows fast data processing with little code. This opens up a range of possibilities, such as the creation of customized web applications and dashboards according to the needs of each project.
Business Area Enhanced: Product & Service Development
Use Case Stage: In Progress
Value Generated:
The metrics used to measure project success were:
- Ease of developing a web application;
- With Dataiku, it is easy to publish a web application and its development is easy, you only need knowledge in Python.
- Ease of user access;
- User access is simple and intuitive because it is possible to publish web applications in workspaces on the platform, making it easier for users to find what they need.
- Time saved in reporting;
- Some labs used to produce reports that summarized analytical data for end users to explore; with this new tool/approach, this is no longer necessary, saving labs time that can reach up to 30 minutes per day. Another important point is that easier access to data also speeds up projects/activities that depend on it, but this is a KPI that will take more time to measure.
Value Brought by Dataiku:
The value that Dataiku brought was the increased democratization of data, since the platform made it possible to create several applications that facilitate the consultation and comparison of historical data. In this regard, it also accelerated the development of projects of this type, since with a common platform for development and access, the time that would have been spent on designing the data infrastructure, providing the development environment, controlling access and publishing, is converted to web application development, since the platform handles all of this in a simple way.
Another point that can be mentioned is the fact that the use of Python through Dataiku enables the development of data visualization tools that were previously discussed, but were not feasible or required a great deal of effort to develop due to the tools previously available.
Value Type:
- Improve customer/employee satisfaction
- Reduce cost
- Save time
- Other - Data accessibility
Value Range: Thousands of $