Sports Statistics App - Advice

WR1 · 2024-09-18T10:01:32.005Z

Hi all,

Relatively new to Bubble and very much in the exploratory/research stage in the development of my app and looking for some advice.

Some context - the app is a Sport Statistics app that will feature both team and player stats for football teams across multiple countries. Consequently, this will be mainly just displaying data to the user as opposed to user creating/editing things/records. MVP for the app will involve around ~30 different leagues. This will mean:

~ 650 teams records - each with 1 row of data & ~70 columns/data points
~ 25000 players records - each with row of data & ~40 columns/data points

This data will need to be updated several times weekly. As well as some other tables like Fixtures across all league (will also require regular updating). As this data is all dynamic, the only way to show it to users will be with Do a Search for/API to an external database. Whilst I know I can minimise searches with hiding data on loads, the concept of the app is to offer end users a wealth of data so the option to display the data will always have to be an option.

I know this achievable in Bubble, but my concern is, with the searches required and the constant updating of data records, will this be feasible in terms of WUs and pricing?

With this in mind I have several questions.

First and foremost, as mentioned, will this be feasible in terms of WUs and pricing due to regular updates required and searches to display the data on the front end?
Would utilising an external database save on WUs as the updating of records would take place outside of Bubble?
Is using an external database viable to display high volumes of data with regards to performance? If so, recommendations on which?
Currently, I’m getting my data in Google Sheets as it’s easy for me to manipulate and manage. Is there any existing way to utilise GSheets as an external database?

Thanks all. Any general advice is appreciated.

rivergalli · 2024-09-18T16:23:20.060Z

We run a data heavy application on Bubble for sports cards and trading card analytics. We run a dedicated instances due to the heavy requirement on processing and data size. But in general there is not a size / growth issue related to bubble other than cost. That said, when you start getting tables with say 50 Million + records you will want to use an external service for indexing (Algolia or Typesense). In general there are no issues storing data in say Supabase and retrieving data from that platform via API calls in a performant way. Additionally, we often use external custom services/code hosted on Replit, if we need certain complex features that require code. All of these things are feasible and we use them in production.

WR1 · 2024-09-19T11:15:06.443Z

Thanks for your @rivergalli.

That’s reassuring to know but for me, the feasibility of this project and the cost, are directly linked.

Interested to know - when you were scaling, were there other methods/integrations you used prior to having your dedicated instance?

rivergalli · 2024-09-19T15:14:24.024Z

@WR1 I completely understand. We load 90K records a day and have real-time calculations running on the back of that which required a lot of CPU/memory. My gut is your data footprint will be smaller but to get an accurate idea of your cost you will want to build a prototype. Testing the ingestion and any calculation processes will give you some idea of the daily cost. There is no question that building w/ code would be cheaper on the infra side.

jagdish_bajaj · 2024-09-19T17:03:12.687Z

What is the revenue model here? It is an important factor to consider since your WU costs will add up and you may want to price access accordingly.

Some users have reported using Xano as a backend for high data volumes - something you can explore, if you haven’t already.

I agree with what has been said earlier - build out a small prototype with a tenth of the data (or less) and see what the costs come up to.

WR1 · 2024-09-20T12:48:12.898Z

Thank you both @rivergalli @jagdish_bajaj

I’ve came to the same conclusion that there are too many hypotheticals at this stage so I need to start building further and see how it goes.

However, my instinct is that using an external backend which such as Xano or Airtable (I’ve been testing both) so I can alleviate all the WU costs attributed to refreshing data as this would all take place outside of system and called via an API which I believe is also cheaper than Do a search for. This is also taking into account the costings of these external platforms .

rivergalli · 2024-09-20T13:22:46.113Z

@jagdish_bajaj I agree with this. You may want to look into Supabase as well. Its a scalalable backend that should support your needs here and they also have a good UI that makes it easy use even with limited SQL skills.