Weekly discussion thread: Oct 20-Oct 26

BI vs data apps :crossed_swords::
BI is forgiving when it comes to data freshness, query latency, and concurrency :hourglass_flowing_sand:. Analysts are interested in looking for insights rather than taking immediate action. This is in stark contrast to data apps, like personalization, customer 360, and fraud detection, where immediate action is needed :timer_clock:. Because of this, these data apps have strict requirements on data freshness, query latency, and supporting high concurrency. Is your company building data apps :thinking:? What challenges are you trying to solve :mount_fuji:?

The Flash Running GIF

For BI, at Rockset we recently migrated to Retool from ReDash (decided to retire their hosted service). Initially we debated between that and Superset/Preset, but decided to use Retool.

The main reason is that it gave us complete freedom to write our SQL in Rockset and just use Query Lambdas, where Superset would require us to use SQLAlchemy, which will prevent us from taking advantage of things like Hints to further improve performance of our queries.

Query Lambdas are also a great option when integrating Rockset for Data Apps and we are looking to build our own C360 app around that - I’ll share more details as soon as I can!

1 Like