3 ms·
Looks interesting. A few suggestions for useful features: Time Series functionality would be handy (Exponentially weighted moving average, centered moving ave
by bigger_cheese 9y ago
Looks interesting.
A few suggestions for useful features:
Time Series functionality would be handy (Exponentially weighted moving average, centered moving average etc).
Aggregations across sliding intervals (by shift / by day/week/month/ Finical year etc). Being able to quickly answer questions like What is the production for week to date, current shift etc. How about previous week/shift etc.
One other comment I will give is I find SQL to be useful as a protection against product lock-in. It's essentially the lowest common denominator language for dealing with data extraction/transformation etc. Our business has been burnt before by having our reporting/BI written in a tool specific language it made migrating away from the product very difficult. Ever since I've very much been of the mindset to do as much using SQL as possible.
- mariusandra 9y agoHi, We already support the most basic sliding intervals: day, week, month, quarter, year. Custom ones would be convenient, I agree. Check the menu when you hover the header of the column "week confirmed_at" here: http://insights-demo.mariusandra.com/explorer?columns=Order.confirmed_at!week%2COrder.user.country.name%2COrder.id!!count&sort=-Order.confirmed_at!day&treeState=Order%2COrder.user%2COrder.user.country&graphTimeFilter=last-365&facetsColumn=Order.user.country.name&facetsCount=6&graphCumulative=false&percentages=false&filter%5B0%5D=Order.confirmed%3Dequals%3Atrue http://insights-demo.mariusandra.com/explorer?columns=Order.... The plan is to add a view called "generated SQL", where you can see and copy/export the raw query that went into generating the data. This way you can always validate the data and get some "insight" into how it's fetched.