Instasights

• 

• 

  Our architecture

• 

• 

• 

• 

Enjoy a working prototype of Instasights by opening the Instasights web app in Google Chrome on your laptop or your smartphone.

Inspiration

Modern companies require modern leadership. Nowadays business insights are often presented in cluttered, bloated and boring dashboards. However, through the likes of TikTok and Instagram, the next generation of leaders is much more accustomed to consuming information in short and fun bite-sized pieces. We aim to bridge this gap.

What it does

Instasights offers an instant overview of the latest important company insights though Instagram-like stories. Our web app includes audio-visual stories that present short summaries of the most important recent insights gathered though our data analysis in the backend. Color coding and animation of the summary texts based on business impact allows for even easier processing. We even offer a daily briefing that includes the most important insights from all topics. We also include summarized relevant business news that can provide context for developments within the company. Additionally, stories can include polls among the companies leadership to evaluate whether one of the automatically found trends requires instant action.

How we built it

We built the web app using Angular in combination with Ionic. The data for the stories is fetched from a Flask server. The server generates the insights from the provided kaggle dataset using kats (including Prophet), pandas and Python. Insights are delivered to the frontend using a custom and generic format that is independent of the used dataset. We use the browser's text-to-speech capabilities to narrate stories and leverage Unsplash to get an image that fits the context.

Challenges we ran into

It was challenging to keep the generated texts short and easily digestible while still providing value. For the frontend it was challenging to implement an Instagram-like player for the stories while animating and voicing over the text. We overcame these challenges through distribution and parallelization among the team members.

Accomplishments that we're proud of

We are proud of being able to present a fully functional prototype that you can try out for yourself. It was also interesting for us to have a deeper look into the provided dataset and to automatically extract insights and weight their business impact.

What we learned

A major learning for the backend team was how to use kats and Prophet for data analysis while our frontend team got to dive deeper into css animations. Furthermore, we learned a lot about working together as a team on a challenging task in the short time frame of a Hackathon.

What's next for Instasights

Our web app is easily extensible with additional insights for further topics by adding more data to the backend. It would also be exciting to enhance the presented data with further context and the ability to dive deeper. Additionally, Instasights could also extended with stories posted by the leaders themselves (in addition to the existing automated data) to allow for them to share information within the company.

Built With

• angular.js
• dedication
• flask
• jupyter
• kits
• python
• rxjs