Dataiku’s end-to-end platform gives our users the ability to turn raw data into business-impacting decisions using advanced predictive analytics. In addition, Dataiku serves a large variety of user personas, technical skills, industries, and use cases. In this blog post, we're going to show you how you can expand Dataiku’s value within your organizations even further by integrating with your BI tool of choice to uplevel from prescriptive to predictive analyses.
I’m Already Using a BI Tool, so Why do I Need Dataiku?
Business teams generally prefer to do their data work on BI tools like Tableau, Power BI, Looker, or Click. Through the advanced visualization capabilities that these tools provide, business teams can quickly create business value from data, but they lack the ability to take their decisions further with advanced predictive analytics. Even teams who do use Dataiku to derive business impacting decisions may reach a gap in communication if they want more advanced visualizations or if more context-driven input is needed in the design of models.
This disconnect is very real in all organizations, but thanks to Dataiku’s external integrations and connectors, we can expand the reach of what is possible. Dataiku has several connectors to the most powerful BI tools on the market to bridge the gap between teams and break down silos of information.
Dataiku’s Visualization Features for Advanced Analytics
Advanced Analytics teams working in Dataiku can create flows to transform raw data and extract business decisions before pushing their data insights into an organization's preferred BI tool.They can also deploy models to an API node for usage within those same tools.
On the other side, BI tool users can visualize the data and predictions they've received from Dataiku to transform it to reflect customer insights and business context, before pushing it back to Dataiku for further work.
Dataiku has a lot of native visualization features, like building charts on top of data sets, performing visual statistics, and bringing those and other kinds of visual insights together inside a dashboard. The focus here, however, is on Dataiku’s third-party integration features. These can be grouped into two categories:
- Those from Dataiku to the third-party solution
- Those from the third-party solution to Dataiku
There is a two-way communication happening here — a two-way data exchange between these two different environments. So the question you should ask yourself is: “From which side does the process get initiated?”
Using Dataiku With Third Party Visualization Solutions: 6 Possible Options
A: From Dataiku to the Third Party Solution
Export to a Proprietary Data Format: A user working on their project from inside Dataiku builds a flow and prepares some data. Once they have a dataset, they can export it as a particular format. More concretely, this consists in downloading a particular data file to their local machine laptop desktop, switching to their visualization solution, and importing that file. They can then continue their work from there. Dataiku currently supports this option for Tableau hyper format, for Qlik sense and QlikView, QBX, as well as for MicroStrategy.
Export Directly From Dataiku to the Third Party Solution: In this case, a user working in Dataiku does not need the additional hop of downloading a dataset inside a proprietary data format to continue working on a third party platform. They just tell Dataiku to send the data instead, directly to the visualization server. We currently support this for Power BI and Tableau.
Pull Data Through a Query or an API: A user who needs additional data from a third-party solution can pull data in through an API. So for any third-party visualization solution which has an API that you can query, perhaps a Python-based API, you can run that Python code inside Dataiku and bring the data in that way. We currently support this for Looker, as we have an integration which uses a predefined look in Looker.
B: From the Third Party Solution to Dataiku
Import a Format: A user working in a visualization solution would create a proprietary data format package from that visualization solution and then switch to their project in Dataiku and import from that data format. You’ll notice that this is exactly the same as the first option, but in reverse! So again, it's it's it's there's no setup is very easy to use. We currently support this for Tableau.
Query a Deployed Model API Service: A user already has a model machine learning model built in Dataiku and has it deployed to an API node in their Dataiku environment, but would like to query that model endpoint from their external visualization solution. So any visualization solution that supports an API client, for example, Python, or or R, or Java, could query that model endpoint.
Using Data Writer’s Public API: You can connect to the entire Dataiku environment from an open public API. Any third party visualization solution that supports creating some kind of plugin or extension or an add on that allows you to connect to third party applications, or an API could take advantage of this capability.
Dataiku and BI tools individually bring so much value to an organization by enabling users to take in a massive amount of data, turn it into something understandable, make data-driven decisions and even look to the future. It's no wonder that the combination of the two has and will continue to increase the business value all that users can create from their data and analytics.