Small Business Insights Dashboard

New actionable insights (the main page of the dashboard), showing a list of categories and insights
Each insight can be accepted or dismissed, which will hide it from this page and move it to the historical insights tab
Once dismissing/accepting all insights, the table has an empty placeholder display
All previously encountered or previous insights are available on the "Historical insights" tab
Loading intermediate state
Initial login screen
When first onboarding, small business tenants have to opt in to share their data with the network in order to gain access to the insights
Original Figma mockup, showing the utilization of NCR's design syst
JSON from our "New Insights" API that is used to display insights in the dashboard

Inspiration

We were inspired when brainstorming ideas for NCR's challenge at HackGT 2020, where we were searching for ways we could utilize our technical expertise to tackle some of the issues that NCR faces on a daily basis.

Our idea grew from realizing the need many (small) businesses face with determining market trends and best products and practices in their industry. While many large companies have access to vast sums of data with which they can inform their business intelligence, small businesses often don’t have that luxury. The ability to crowdsource data could benefit small businesses with making product, staffing, and promotional decisions based on business growth models, allowing them to better compete with larger companies.

Our project addressed several of the top problems for NCR and its customers: namely,

How NCR can help small businesses thrive
How NCR can help reimagine commerce and disrupt aspects like product discovery, e-commerce, payments, and supply chain
How NCR can help merchants monitor and optimize their performance
How NCR can help merchants reduce their labor costs

What it does

The project aggregates data across all (NCR businesses) and categorizes them and their products into categories. The app then calculates the monthly growth for the business and recommends products to add to a businesses catalog, products to remove, when to increase staff, and possible promotions, based on data of all other businesses in the same category.

Through the dashboard, businesses can track recommendations for the current month as well as all past recommendations and their choices on those recommendations (accepted or rejected).

How we built it

We built Insight with a react frontend and flask backend. The frontend was designed in Figma and implements NCR’s design system using hand-written CSS with React.

The backend utilizes Python and Flask to create routes that the frontend interacts with. The backend API is containerized and orchestrated onto a local server. The backend also has NCR’s Business Service Platform’s Catalog Service implemented, storing product data.

Challenges we ran into

We struggled with the NCR Catalog Service API, finding a use case for it because it required the uploading of our own data. While we did end up successfully uploading our data into the NCR Business Services Platform and successfully connected to it, it was easier to perform aggregation and analysis locally. That being said, in a scaled system where the catalog service provides product data for thousands of businesses, it would make more sense to use it.

Accomplishments that we're proud of

We are proud that we were able to ideate, mockup, and create a product from our mockup within around 15 hours. We are proud of our ability to parallelize tasks and split them up between our team, allowing for an infeasible task to be completed. We are also proud of our research, both in finding data that suited our use case, as well as the research of models that represent business growth when designing our algorithm.

What we learned

Working on this project allowed us to learn a variety of new skills that we wouldn’t have been able to learn otherwise, such as:

Creating high-fidelity UI mockups from and existing design system and then translating those mockups into actual frontend code using CSS-in-JS
Applying business growth models to create actionable insights from data
Utilizing NCR’s Business Services Platform API’s to build our application

What's next for Small Business Insights Dashboard

In general, we envision being able to take the next step and make the Insights dashboard a fully-fledged product that small businesses can benefit from. Some of the ideas we have are:

Scale our project to operate at the scale of NCR, allowing small businesses to benefit from the wealth of data and the insights it produces
Employ unsupervised machine learning to generate additional insights based off of more complex patterns
Create more insights based off of BSP or other NCR services
Expand the use cases of the tool to include other businesses that would benefit from sales or product-based insights, such as chain restaurants