Ask HN: Modern BI stack 2022 for low entry and cheap license dashboards | Hacker News Reader

Ask HN: Modern BI stack 2022 for low entry and cheap license dashboards

7 pointscaliforniacream8 comments

Hi all, I'm looking to get into niche dashboards for my special interest (cycling adjacent and adrenaline junkie). This world has tons of users, but no 'whales'.

I simply want to host a dashboard and allow people to access it for 10-15 dollars a month while making a bit of meat for myself (to allow full-time work on the project). I considered PowerBI, but learnt that the user account license would run 10 dollars a pop, per month. So say I scale to 10k users, almost all of my money would be going straight to the MS.

What are the best modern ways to do this? Looking around, it seems I can use an Azure SQL server or Amazon dynamoDB to host the database and get good reliability and concurrency somewhat cheaply. How can I do the same with the dashboard itself? (if I was doing time series data, an Amazon managed Grafana looks like a winner, but I'm doing a simple dashboard with little time series data).

Thanks!

If you are technical, I would recommend looking at the open-source Python ecosystem to create reports and dashboards. It has amazing analysis and visualization libraries (like Altair and Plotly).

We are building an open-source library to create data front-ends like reports and dashboards using Python (https://github.com/datapane/datapane). You could programmatically create a dashboard for each user and save it as HTML to something like S3. As the dashboards can be pre-rendered and static, they can scaled really heavily if you need.

+1 on this

My employer https://SigmaComputing.com offers public, interactive iframe embeds of dashboards for zero marginal cost. Here is an example of MMO players making a dashboard https://redd.it/lrn303. I would say Sigma is the best tool on the market for telling stories with data, but not as pretty as Tableau.

CubeJS is a great piece of software for dynamically generating reports from SQL or other data sources. Queries are either written in their JSON query format or in SQL. But it's headless so you have to build a frontend yourself.

Google Data Studio is free and easy to use. As for database I recommend BigQuery which is well integrated with Data Studio. Don’t forget to post the dashboard on HN, I’d love to see it!

AWS has quicksight but it's pricey.

You're probably better off just building some visualisations manually using a charting library.

Yep. Any ideas on how to serve it to 1000's concurrently? Also, is there anyway to be sure that most/all of the people who have payed can view the content?

With PowerBI, MS handles that - what server/library can I use to know if people can access? Sorry if this is a dumb question, I'm new to the space!

You can probably put a react app together with cognito, a lambda and API gateway. Maybe chart.js or some such for the visualisation?

Reply on news.ycombinator.com