Graph Snippets
How might we reduce the time spent on data preparation to help data scientists focus on their primary goal of a data analysis expert?
My Role
Timeline
Impact
I was responsible for brainstorming, research, ideation, information architecture, user flows, visual design, interaction design and prototype. Once it was rolled out to development, I collaborated with the development team to ensure that the development and design goals align.
2 Months
We reduced the learning curve around the modeler and the time taken to build and use data pipelines by introducing easy to use, modular and customisable graph snippets
The Overview
Data sciences is one of the most sought after professions in the industry. Most industries have huge data lakes of which they don't know what to do. Here is where Business analysts come into the picture to make sense of unstructured data. They are able to identify trends and patterns that can help companies venture into new domains and find new business opportunities. In order to do all that, some companies use the SAP Data hub to facilitate their findings.
What is SAP Data hub?
SAP Data Hub is software that enables organisations to manage and govern the flow of data from a variety of sources across the enterprise. SAP Data Hub allows companies to build a pipeline of data from various sources - which can include ERP, data warehouses, big data lakes, and more - that can then be used in new data-driven applications.
A single management layer essentially treats all data the same, regardless of the source, which means the data can be governed, managed and subsequently used in applications.
"In simple words"
Or at least that's how I explained it to myself and others...
Imagine today you are incharge of the kitchen (yay!) and you are cooking! You decide, "I want to make a sandwich today." What do you do?
In this case, the Data hub is essentially a kitchen where you ingest data (ingredients), prepare raw data which can be consumed. Now this is a simple example. Now envision you running a restaurant and you are the owner. You would need to assign duties, give certain permissions to individuals, form new connections with your vegetable vendors and maybe monitor overall performance of the tasks. That is what SAP Data hub is.
This is the SAP Data hub landing page.
What is Modeler?
The SAP Data Hub Modeler tool is based on the SAP Pipeline Engine that uses a flow - based programming paradigm to create data processing pipelines (graphs). Big Data applications require advanced data ingestion and transformation capabilities.
"In simple words"
The Modeler is essentially a manager at your restaurant. The modeler would help create pipelines or a flow of work that would help manage your tasks just the way a manager would manage the staff and tasks.
This is the Modeler tool
The Ask
As the sole designer taking lead, I was asked to design an enhancement into the product. The main stakeholders wanted to create a library of templates which people could reuse. This was not a scalable solution and had a lot of limitations. I pushed for thinking in products to make sure that we build the right features for the right audience and tackle the real problems that the users are actually facing or currently have. The real question to be addressed was:
How might we reduce the time spent on data preparation and
simplify the process of building a data pipeline?
Introducing : Graph Snippets
After a lot of back and forth between my product owner and the developers, and after many iterations, we came to introduce this feature called Graph Snippets. From an existing graph, one could create presets or templates which would help our users and their teammates to speed up their data preparation process.
After a lot of back and forth between my product owner and the developers and after many iterations we came to introduce this feature called Graph Snippets. From an existing graph, one could create presets or templates which would help our users and their team mates to speed up their data preparation process.
Import your Snippet
Change the way you create graphs. Instead of creating a new graph every time from scratch, select default snippets shipped with product or select customised snippets tailored to your needs.
Use a wizard format to glide through essential properties only as the system pre-fills others properties by default.
User friendly Wizard
Create your Snippet
As an experienced user, you probably know what's best for you and your team. You can customise the default snippets or create your own, tailored to your workflow.
Essentials Only
The idea is to select only the essential properties that would require specific input from the user
Master Tab
Freely select operator of your choosing to work with. You can also re-arrange the sequence of your pipeline as per your choice, giving you complete freedom and control.
Detail Tab
With the help of check boxes, you can cherry pick essential properties to be displayed in the snippet.
How did I get here?
The WH questions prove to be a great starting point in order to understand and decode the problem at hand. The answers to these six questions reveal the context of use – the people, their tasks, their tools and technology, and the environment in which they will be using the product you’re designing.
In order to understand the motive behind this said feature I had to prioritise the why aspect, then the how and then the what following Simon Sinek's Golden Circle. This helped us steer away from a feature first approach and more towards the humans first.
We wanted to make a user friendly tool which helps business users with data preparation
WHY
(HOVER)
We wanted to help naive users onboard the tool with ease
HOW
(HOVER)
With the help of a stencil or template, we could simplify the tool complexity
WHAT
(HOVER)
Challenges and Hurdles
Undertaking this task of designing the snippets had its own set of challenges. I had to conduct my own research to understand the problems and solve these challenges.
-
SAP Data hub is a fairly new product launched in 2017. Due to this, I did not have access to any user data, persona, feedback, suggestions etc. There was no user database to get prerequisites. My only source of information was my Product Owner who knew of the customer and business goals.
-
There were no existing use cases or scenarios to draw inspiration from. To build these templates I needed to know commonly used graphs or combination of operators used in a pipeline. Even a real life scenario where how the data hub was used was not available.
-
During this project, the entire tool was under going a design revamp following "SAP Fiori 3" guidelines. As Modeler is a big tool, the Fiori 3 alignment for Modeler was put on hold. So I had to adapt my concepts built for Fiori 3 to the Fiori 2 version of Modeler.
-
The requirements itself were not properly defined. Nobody in the team knew what exactly was expected so we had to define the problem and deliver the product.
Investigating for Information
We were designing for Business Analysts and Data Scientists. I had no access to real users, so I set up one-on-one interview with my Product Owner and the Product Marketing expert to extract information. Doing this helped me gain critical information like:
From Product Owner
-
Drawing on the PO's knowledge of the market and customer base so that I don't miss the business goals
-
Getting suggestions and input on features that would help increase the success of the product
-
Understanding performance indicators to measure Product development and Success
From Product Marketing Expert
-
How Data hub is used as a product
-
Understanding real life scenarios where the Modeler is used
-
Common use cases to explore
From Developers
-
Technical understanding of the tool and its components
-
Actual pipelines to see how graphs really work
-
Technical limitations
I still needed to interact with a business analyst to connect at an emotional level; to really understand the pains in their workflow. Luckily, one of my friends happened to be an Associate Business Analyst. I had a one-on-one conversation with him to understand his perspective of the job.
Observation and Insights
Talking to my friend gave me several insights about what was critical as a business analyst. Some of the main observations are as follows:
-
Business Analysts spend majority of their time preparing data. Their actual job comes after this but preparing raw data is equally important.
-
Preparing data is a time consuming process and some tools make it really complex to use
-
Analysts use different programming languages in their respective workflow
The insights I drew from these observations are:
-
Business Analysts need to speed up their process
-
They need something simple and easy to use
-
Approach needs to be modular to accommodate versatility
User Persona
After understanding the mental model of who we were designing for, we could break down the personas into two types: The Rookie and The Expert. Keeping these two in mind, I developed the following personas
The Rookie
The Expert
Ideation and Concept Generation
Once I had the relevant information regarding my problem, I started jotting down my thoughts, questions and ideas. I resorted to WH questions again, to frame my thought process.
Questions are the basic tools in any designer's toolkit.
Decoding the template
Early ideas and concepts
Low Fidelity wireframes
Wireframes and Prototypes
Conclusion
After gruelling hours of back and forth between the stakeholders and after multiple iterations, we finally produced the "Graph Snippets." It also underwent different nomenclatures and naming conventions. From design stage to production was very taxing but it taught me a lot regarding tiny details and contingencies.
The entire production from design to development took about a span of 5 months. It finally got released to the main branch of the product and it is available for customers to consume! The product is live and running.
Take Aways
ASK QUESTIONS
Ask loads of questions like you're a child. We often assume that when given a brief, everything else has been thought through and all you need to do is design. That is kind of irresponsible as a designer because turns out we are in the same boat. Nobody exactly knows what is expected and which is why asking the right questions help you create a path towards the requirements.
KILL YOUR DARLINGS
We often fall in love with our first designs. In our minds, that creation triumphs everything else and seems to solve all problems. We overlook the shortcomings of our creations. It is important to let go of these ideas to truly craft for; not yourself but the humans you are designing for.
PEN AND PAPER IS KING
I have always found it handy to draw on paper first as your thoughts are still very raw and gives you flexibility to make mistakes and record those mistakes. You never know when an idea can strike! Just translate your thoughts into rough sketches and watch your mind bloom.
DEVS ARE YOUR FRIENDS
Listen to your users and stakeholders. As a designer you need to know what to compromise upon for the longevity of the product. It is important to get feedback from your developers and understand the technical constraints. On the brighter side, these constraints become your sandbox and force you to think out of the box to overcome the problem.
