Hi everyone,
I do not work in a business unit or in BI, but I'm like an analyst for communites, cities, regions, and every type of local administration.
So I often have to analyse large quantities of datas for them, to help them understand their territory, the needs, the priorities for the public policies, etc. Thanks to the open datas, we now have plenty of datas to analyse, aggregate, compare, etc. but it's also time consuming and complex.
That's why, instead of producing every time the same indicators, maps, charts… for different territory and administration, I started to identify and save every relevant open data I need, organize it in a "BI way" (fact table, i.e. population, and dimensions, i.e. years of census, category of age, geography, range of incomes, etc.)
My core analyses can be with 30/40 different databases, but in a perfect world, by the end, I'll probably have 100+ fact tables integrated and reported (I store and manage those database in postgresql).
What would be very usefull is primary to filter all that tables by geography (i.e. selecting a city and some others for comparison) and edit sort of a big data booklet, PDF and/or interactive, that I can use for my analysis. Even though there will be differences from one territory or administration from another, at least some basics will be the same so a « copy paste » report would be a great starting point anyway.
I started to work with Pentaho CE (creating cubes with schema workbench and Mondrian, analysing with Saiku and after that, producing reports), but had some difficulties with my cubes, and found some limits with the old fashioned and not quite-so-good-looking reports of Pentaho, compared to what Qlik can produce easily. And the fact that I don't need to create cube was appealing. Also, a free solution, at least for testing, concepting, learning and basic reporting, was key.
But as I discover more and more Qlik solutions (I'm using right now Qlik Sense Desktop as a way to initiate me), I'm not quite sure right now that it suits my needs, essentially the possibility to create filtered reports for a large quantities of fact tables.
What do you think of that ?
I hope I've been clear for you. Else, do not hesitate to ask for precisions.
Thanks !
