Is there a way to select a staff and set all voices to the same number?

As mentioned, it's a mess precisely because it's from MIDI, which was never meant for conveying notation. So, unfortunately, turning that into readable notation is probably going to take more work than simply entering the music normally. Especially if you're not already an expert MsueScore user, since entering music normally is way easier to learn than solving complex problems like this.

But that said, Implode will still do what it is designed to do. you simply need to do it one staff at a time. So first select the contents of the top staff, then run Implode, then do the same for the bottom.

What you'll find is, anywhere it actually possible to combine the voices - anywhere both voices have exactly the same rhythm - they'll be combined. But it won't make much difference here, but the music as shown here is not possible to represent in only one voice. The rhythms are too different. In order to notate on one voice, you'd have to be careful to only play it using a single voice. So, no starting one note before letting go of the previous. Thats the thing that requires multiple voices. Not because of some oscure quirk in MsueScor,e but because that's how music notation works. Any time you start a note while another is still sounding, that requires multiple voices. And that happens a lot here.

Midi is generally not a very useful format when you are dealing with notation. Indeed, midi just does not contain notation information and in particular it has no concept of notated durations - minim, crotchet, quaver (or half, quarter, 8th for US readers) etc. It just records when a note starts and when it ends When midi is imported to a notation program some AI has to be applied to work out for example which note duration value to use for a note that starts at x tics and ends at x + y tics, - including things like "is that a staccato crotchet or a quaver, or perhaps a dotted quaver?" and if it is a midi rendition of a human performance, it has to take into account possible inaccuracy (or to be kinder, variation) in the midi timings themselves caused by a human player using subtle rubato, or just plain mistiming things. This means that what should be chords of identical length notes may in fact end up requiring the use of multiple voices to accommodate slightly different start and end times.

Your original messy version looks like it is a midi import. You will probably get better results if you process the midi before trying to import it to MuseScore, In particular look at setting up instrument channels to separate piano left and right hands and quantizing note lengths. MuseScore has some limited facilities for processing midi during import, but a dedicated midi editor is probably going to do a much more useful job of it.