Data Science
Machine learning & data science for beginners and experts alike.
Yin and yang, spaghetti and meatballs, ham and cheese. All these combinations work in perfect harmony. Now the challenge is: “How can data scientists and analysts also work in perfect harmony?”
Data scientists, like their name suggests, have amazing insight into the world of data: processing it, cleaning it, ingesting it, analyzing it, and predicting what comes next. In working with many data scientists, they have helped enable better clustering, more accurate forecasting, and better insights into what issues are arising in the data. They have so much knowledge in many coding languages like R, Python, SQL, etc. They tend to push the envelope on how data can be consumed, always learning new ways to look at data, and processing it.
On the other side, analysts learn their business and understand how operations and small adjustments can heavily affect their business. They know that a small adjustment to assortment changes how much labor is needed, what a promotion could do to the movement of units, how signing a new contract with a supplier can be a big win for supply chain, or how a new policy impacts associates. They know how to find the intricacies of the business and how to pull levers and buttons to make an impact.
Now introducing the newest Yin and Yang combination: DATA SCIENTISTS AND ANALYSTS!
The combination of business knowledge and data analytics should be the goal of this yin and yang harmony. Alteryx empowers this combination to be so much stronger!
The main question you probably have is: “How can Alteryx blend these two areas together?” I think there are several ways Alteryx can help bridge this gap:
- Alteryx Analytic Apps
- Code-Friendly Tools within Alteryx
- SDKs within Alteryx
- Eventually Assisted Modeling within Alteryx (I am so excited for this)
All four of these methods can help increase collaboration between data scientists and analysts. Each one promotes a different way the two can partner and collaborate on different issues business faces. Each also provides different ways to come to a solution allowing flexibility and freedom within solutions.
The Alteryx Help gives insight into what an analytic app is: “An analytic app is a workflow with a user interface. Create an analytic app to enable the app user to execute a workflow using their own data and parameter without having to build the workflow.” Based on this definition, Alteryx Analytic Apps are a great way to allow analysts to interact with data scientist-built models and adjust them to different specifications without ever needing to code. Data scientists could then enable other analysts to create and deploy models all without ever coding. The data scientists could also employ different checks to make sure the model passes standard tests for good models within each developed app. Through this collaboration, analysts and data scientists can both push business and analytics to higher levels.
Through my own experience, I have seen the power Alteryx has in being both code-free and code-friendly. It allows data scientists to develop R or Python code and help a data analyst implement that code right into their existing Alteryx workflows. Throughout several projects, I have partnered with a data scientist to build models that are then implemented into a workflow so data and parameters can be a bit more dynamic and also allow us to automate and schedule the workflows. The ability to drop code straight into a workflow is invaluable in the automation process and gives freedom to both analysts and data scientists to continue to push envelopes and develop even more.
SDKs with Alteryx can empower a data scientist to build models and modelling into a prebuilt and customized tool for other analysts to use. It allows companies to control the models for different variables but allows the analysts to input what data the model is built upon. I have just recently started with SDKs and I have found so much power in them. They can incorporate so much work into a tool that you can send out to coworkers to use with ease. That can allow a data scientist to wrap their models into simplistic user interface tools for analysts to incorporate right into their workflows.
The final piece that will definitely help bridge data scientists and data analysts is assisted modeling within Alteryx. This is one of the many exciting things previewed in Alteryx within the last year! You can find a demonstration here. Assisted modeling will be a game-changer for data scientists and analysts to partner on understanding models and building models for new projects. Data scientists can help empower analysts to make the right decisions for models and assist them in checking their models all without having analysts code again.
Alteryx has so much power and it truly amazes me how much it can bring analysts together with data scientists, simplify their collaboration, and increase the effectiveness across all parties. With the strides forward Alteryx is making (full Python library, assisted modeling), I cannot wait to see how they continue to improve the harmonization of the newest yin and yang: Data Scientists and Analysts.
