I used the numbers of a German casino. I ran the entire data, not just few numbers, every possible set of 6 and more numbers to find if there were returning groups. I found that it was good random, so what was returning wouldn't have paid. The next day the results changed, so you had nothing, it all looked very random and this often was said too.
I also did similar with pick 3, not finding anything that I wanted to really play.
I suppose that some hope to turn pick 3 into a deck of cards. I had remarks on my choices in posts years ago. These have this wrong idea that once combinations of your set were drawn, you could eliminate them. This is wrong. There is no reason why a combination would not be drawn again. I also think that those remarks are poison in lottery forums.
I also analyzed followers, digits followers in different ways. There was nothing useful. That doesn't mean that in some state the random events don't have a repeating abnormality, or something recognizable that you might want to play.
If you only look at numbers for lotto, not looking at combinations, then it is doable with Excel. What I did there, Excel cannot do.
With boxed combinations you have more luck, those are like wheels of 6 straight combinations. The straight combinations have a stronger leverage, better payout. The boxed doubles are more interesting.
I don't eliminate, digits can repeat, it can be a choice to not play those. I coded repeat charts for digits with more detail, and they were fine. I didn't find any rule, it always was guesswork.